---
title: "AI tools: CLI agents, editors & model clients · Undominated.ai"
canonical: https://undominated.ai/tools/
description: "Compare 112 AI tools by workflow, model access, local support and licence. Current sources, clear trade-offs and side-by-side comparison."
---

# AI tools: CLI agents, editors & model clients · Undominated.ai

> Compare 112 AI tools by workflow, model access, local support and licence. Current sources, clear trade-offs and side-by-side comparison.

THE AI TOOLKIT / INDEPENDENT FIELD GUIDE

# AI Tools. Choose with evidence.

Compare terminal agents, coding editors and model clients. Find a tool that fits how you work, with the reasons, limitations and sources beside it.

 [Explore the catalogue ↓](#catalogue)[How we make a recommendation ↗](#method)

 RESEARCH SNAPSHOT tools in this index 112
 profiles checked against official sources 109
 distinct supporting source URLs 364

Reviewed 2026-09-21 [Sources, scope & known gaps ↗](#sources)

 **There is no defensible universal #1 here.** Model choice, permissions, workflow and cost change the answer. This guide recommends documented fits; it does not turn feature counts into coding-quality scores.

01 / FIND YOUR FIT

## Where do you want to work?

Choose a starting point. Refine the requirements below.

 All tools 112 **Explore everything** Browse coding agents, editors, local clients and hosted assistants. Selected → CLI / TUI 70 **Work in a terminal** Command-line tools with a documented terminal interface. Explore matches → Flexible CLI 51 **Choose your provider** Terminal tools that document more than one model provider. Check endpoint limits. Explore matches → Editors 42 **Stay in your editor** A coding editor or a documented integration with one. Explore matches → Local models 53 **Connect local models** Tools with a documented local model connection. Hardware and feature limits still apply. Explore matches → Desktop 34 **Use a desktop app** A graphical desktop interface; this alone does not make a tool an IDE. Explore matches →

CHOOSE BY JOB

## A considered place to start.

Editorial starting points for different jobs. Not benchmark winners or paid placements.

These choices favour a clear match to the job described, current availability, and documented setup. The alternatives change the trade-off. Use the full catalogue to test our choices against your own requirements.

A terminal agent with provider choice

### [OpenCode ↗](#opencode)

Start here when configuring the model provider is central to your workflow. Its documented terminal client supports provider configuration without tying the whole workflow to one model family.

**Weigh up:** An endpoint being configurable does not make every model equally capable or every feature compatible.

 **Compare with**

[goose](#goose) — you also want a desktop automation interface.

[Pi](#pi) — you prefer to assemble a smaller core with extensions.

[Inspect the evidence · 2026-09-21 ↗](#opencode)

Work specifically with Claude

### [Claude Code ↗](#claude-code)

A first-party Claude workflow with documented terminal skills, MCP integrations and delegated work. It is a relevant starting point when you have already chosen that model ecosystem.

**Weigh up:** The supported model path is Claude. A subscription and direct API usage can have different limits and billing.

 **Compare with**

[OpenCode](#opencode) — switching model providers is a requirement.

[Cline](#cline) — you want to compare an editor-centred workflow.

[Inspect the evidence · 2026-09-21 ↗](#claude-code)

Work with Codex and OpenAI accounts

### [Codex ↗](#codex)

Start with the first-party account and coding workflow, while retaining an open-source CLI. Keep its local terminal configuration distinct from the desktop and cloud products.

**Weigh up:** CLI licensing and configurable local providers do not automatically apply to every Codex interface.

 **Compare with**

[OpenCode](#opencode) — provider-neutral terminal configuration matters most.

[Aider](#aider) — you prefer a Git-centred terminal editing loop.

[Inspect the evidence · 2026-09-21 ↗](#codex)

Build your own agent workflow

### [Pi ↗](#pi)

A deliberately small terminal core with documented extension and programmatic interfaces. Choose it when controlling the harness is part of the work.

**Weigh up:** Native MCP and subagents are not bundled into the core. Extensions and permission choices become your responsibility.

 **Compare with**

[OpenCode](#opencode) — you want a more integrated terminal agent.

[goose](#goose) — MCP-based automation is the starting requirement.

[Inspect the evidence · 2026-09-21 ↗](#pi)

Put model calls into shell pipelines

### [LLM ↗](#llm)

The command-line and plugin workflow is built around reusable model calls. It is a focused starting point for scripts and terminal data work.

**Weigh up:** A CLI for model calls is not automatically a full repository-editing agent.

 **Compare with**

[AIChat](#aichat) — you want conversational shell and agent workflows.

[Mods](#mods) — a compact pipe-oriented interface fits your task.

[Inspect the evidence · 2026-09-21 ↗](#llm)

Keep the editor at the centre

### [Zed ↗](#zed)

An editor with documented model connections and external-agent integration. Start here when you want the editor to remain your primary workspace.

**Weigh up:** An external agent keeps its own capabilities, authentication and billing. Editor integration does not make those agents equivalent.

 **Compare with**

[Cursor](#cursor) — you prefer an integrated commercial coding workspace.

[Cline](#cline) — you want to keep a supported existing editor.

[Inspect the evidence · 2026-09-21 ↗](#zed)

Start with a local model desktop app

### [LM Studio ↗](#lm-studio)

A graphical starting point for downloading, loading and talking to local models, with a documented local serving workflow.

**Weigh up:** Model size and hardware determine what you can run. LM Studio and its Bionic agent are separate products.

 **Compare with**

[Jan Desktop](#jan) — you want to compare another local-first client.

[GPT4All](#gpt4all) — desktop chat with local documents is your focus.

[Inspect the evidence · 2026-09-21 ↗](#lm-studio)

Connect tools for broader automation

### [goose ↗](#goose)

Documented terminal and desktop entry points with MCP-based tool integrations make it a useful automation starting point beyond a single editor.

**Weigh up:** Model selection, permissions and interface-specific capabilities still need evaluation for the tasks you automate.

 **Compare with**

[Hermes Agent](#hermes-agent) — a persistent assistant workflow is your priority.

[OpenCode](#opencode) — repository coding is the main job.

[Inspect the evidence · 2026-09-21 ↗](#goose)

02 / THE CATALOGUE

## Make an informed shortlist.

Alphabetical order. Every match is shown. Select up to three tools to compare.

 Search tools Tool type Every type Coding agent Coding editor Model client AI assistant Framework / platform Identity under review Availability All records Current tools Renamed / retired / unmaintained Not fully verified
 Must have — only documented support qualifies Terminal / CLI Editor integration Desktop app Web interface Open source Multiple model providers Local model connection MCP Agent skills Subagents Hooks / extensions
 Install instructions for Choose your computer’s OS Linux macOS Windows Choose the computer you will install on
 Copy this view ↗

**112** of 112 records

 Reset filters

AI tools in alphabetical order. Expand each profile for selection reasons, limitations and sources.

| Tool / purpose | Work surfaces | Models | Licence / access | Compare |
| --- | --- | --- | --- | --- |
| [AdaL](#adal) Coding agent · SylphAI Coding and automation agent with terminal operation, worker orchestration and configurable model access. | CLI | Multiple providers Local connection documented | Licence not established Included credits and paid/BYOK options | Compare AdaL Add |
| Reasons, limits, setup & sources 8 sources · 2026-09-21 AdaL Product mark · Profile reviewed 2026-09-21 Why consider AdaL Coding tasks that delegate to specialized workers. Combining skills, hooks and local or hosted model connections. What to weigh up Local-model support is preview and excludes some capabilities such as image input. Included credits, provider subscriptions and BYOK have different billing paths. Models & access Multiple hosted providers, third-party subscription access and local Ollama models are documented; local support is preview. Included credits and paid/BYOK options Plan credits and your selected model-access option determine costs; check the current Pricing & Usage terms. [Check current plans and usage terms ↗](https://docs.sylph.ai/getting-started/models/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://adalagent.ai/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://adalagent.ai/) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers [Multiple providers ↗](https://docs.sylph.ai/getting-started/models/) Local model connection [Documented ↗](https://docs.sylph.ai/features/local-models/) MCP [Documented ↗](https://docs.sylph.ai/features/mcp-support-proposed/) Agent skills [Documented ↗](https://docs.sylph.ai/features/plugins-and-skills/) Subagents [Documented ↗](https://adalagent.ai/) Hooks / extensions [Documented ↗](https://docs.sylph.ai/features/lifecycle-hooks/) Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://adalagent.ai/) Skills installation target: .adal/skills/ · Global: ~/.adal/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Adal Home ↗](https://adalagent.ai/) official docs · checked 2026-09-21 [Adal Models ↗](https://docs.sylph.ai/getting-started/models/) official docs · checked 2026-09-21 [Adal Local ↗](https://docs.sylph.ai/features/local-models/) official docs · checked 2026-09-21 [Adal Hooks ↗](https://docs.sylph.ai/features/lifecycle-hooks/) official docs · checked 2026-09-21 [Adal Mcp ↗](https://docs.sylph.ai/features/mcp-support-proposed/) official docs · checked 2026-09-21 [Adal Skills ↗](https://docs.sylph.ai/features/plugins-and-skills/) official docs · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [AIChat](#aichat) AI assistant · AIChat contributors Terminal assistant combining chat, shell help, RAG, tool calls and model-provider selection. | CLI | Multiple providers Local connection documented | [MIT OR Apache-2.0 ↗](https://github.com/sigoden/aichat) BYOK or locally served models | Compare AIChat Add |
| Reasons, limits, setup & sources 1 sources · 2026-09-21 AIChat Neutral lettermark; product logo not established · Profile reviewed 2026-09-21 Why consider AIChat Combine shell assistance, document retrieval and tool use in one CLI. Use hosted and local models without switching the interface. What to weigh up RAG and tool integrations require configuration. Role/agent templates are not evidence of a complete autonomous coding workflow. Models & access Supports hosted providers, Ollama and compatible API endpoints through one configurable client. BYOK or locally served models Configure a hosted model provider using your own account, or connect to Ollama or another compatible local service. Hosted usage follows the provider's rates and limits; local operation uses your own hardware. The client does not establish one bundled inference price across providers. [Check current plans and usage terms ↗](https://github.com/sigoden/aichat) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/sigoden/aichat) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/sigoden/aichat) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/sigoden/aichat) Multiple model providers [Multiple providers ↗](https://github.com/sigoden/aichat) Local model connection [Documented ↗](https://github.com/sigoden/aichat) MCP [Documented ↗](https://github.com/sigoden/aichat) Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/sigoden/aichat) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [AIChat README ↗](https://github.com/sigoden/aichat) official repository · checked 2026-09-21 | | | | |
| [Aider](#aider) Coding agent · Aider contributors Terminal pair-programming tool centered on editing an existing Git repository, codebase maps and reviewable commits. | CLI | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/Aider-AI/aider/blob/main/LICENSE.txt) Provider API usage or local compute | Compare Aider Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Aider Product mark · Profile reviewed 2026-09-21 Why consider Aider Make scoped code changes with repository maps and Git integration. Choose a preferred hosted or local model instead of a vendor-managed catalog. What to weigh up Model compatibility and edit reliability vary; read the model-specific warnings. Skills/MCP/subagent parity with larger agent platforms was not established. Models & access Supports many hosted providers and local endpoints such as Ollama and LM Studio. Model-specific warnings and settings are documented. Provider API usage or local compute Connect your own cloud-provider credentials or run a compatible local model. Some providers offer limited free API access; others meter model usage. Aider's client license does not include a hosted model allowance, and self-hosting still requires your own compute resources. [Check current plans and usage terms ↗](https://aider.chat/docs/llms.html) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://aider.chat/docs/install.html) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/Aider-AI/aider) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/Aider-AI/aider/blob/main/LICENSE.txt) Multiple model providers [Multiple providers ↗](https://github.com/Aider-AI/aider) Local model connection [Documented ↗](https://github.com/Aider-AI/aider) MCP Unknown Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/Aider-AI/aider) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Aider README ↗](https://github.com/Aider-AI/aider) official repository · checked 2026-09-21 [Official documentation ↗](https://aider.chat/docs/) official docs · checked 2026-09-21 [Apache license ↗](https://github.com/Aider-AI/aider/blob/main/LICENSE.txt) official repository · checked 2026-09-21 [Aider installation ↗](https://aider.chat/docs/install.html) official docs · checked 2026-09-21 [Aider model connections, free API routes and local models ↗](https://aider.chat/docs/llms.html) official docs · checked 2026-09-21 | | | | |
| [AiderDesk](#aider-desk) Coding agent · HOTOVO Desktop orchestration workspace with reviewable edits, isolated tasks and configurable model providers. | desktop | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/hotovo/aider-desk/blob/main/LICENSE) Open-source client; model costs separate | Compare AiderDesk Add |
| Reasons, limits, setup & sources 2 sources · 2026-09-21 AiderDesk Product mark · Profile reviewed 2026-09-21 Why consider AiderDesk Review context, token use and proposed changes in a dedicated app. Delegate tasks to specialized subagents while keeping an existing editor. What to weigh up Model providers and approval policies still need configuration. AiderDesk is a separate application, not the Aider terminal interface. Models & access Cloud providers and Ollama are documented. This is an orchestration app distinct from Aider CLI. Open-source client; model costs separate Provider usage or local hardware supplies inference. [Check current plans and usage terms ↗](https://github.com/hotovo/aider-desk) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/hotovo/aider-desk/releases) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://raw.githubusercontent.com/hotovo/aider-desk/main/README.md) Web interface Unknown Open source [Open source ↗](https://raw.githubusercontent.com/hotovo/aider-desk/main/README.md) Scope: desktop Multiple model providers [Multiple providers ↗](https://raw.githubusercontent.com/hotovo/aider-desk/main/README.md) Scope: desktop Local model connection [Documented ↗](https://raw.githubusercontent.com/hotovo/aider-desk/main/README.md) Scope: desktop MCP [Documented ↗](https://raw.githubusercontent.com/hotovo/aider-desk/main/README.md) Scope: desktop Agent skills [Documented ↗](https://raw.githubusercontent.com/hotovo/aider-desk/main/README.md) Scope: desktop Subagents [Documented ↗](https://raw.githubusercontent.com/hotovo/aider-desk/main/README.md) Scope: desktop Hooks / extensions [Documented ↗](https://raw.githubusercontent.com/hotovo/aider-desk/main/README.md) Scope: desktop Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://raw.githubusercontent.com/hotovo/aider-desk/main/README.md) Skills installation target: .aider-desk/skills/ · Global: ~/.aider-desk/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [AiderDesk official README ↗](https://raw.githubusercontent.com/hotovo/aider-desk/main/README.md) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Amp](#amp) Coding agent · Amp Coding agent spanning local terminal sessions and cloud workspaces, with configurable model routing and shared threads. | CLI IDE desktop web | Multiple providers Local connection unknown | Licence not established BYOK, subscriptions or usage billing | Compare Amp Add |
| Reasons, limits, setup & sources 9 sources · 2026-09-21 Amp Product mark · Profile reviewed 2026-09-21 Why consider Amp Move a task between local CLI, cloud Orbs and native apps. Use shared threads and workspace model routing. What to weigh up Model routing and gateway access have plan restrictions. Cloud machine usage and agent usage have separate allowances. Models & access Uses multiple model families. BYOK and linked subscriptions are supported; compatible gateway connections remain early access on selected paid tiers. Local-model operation was not verified. BYOK, subscriptions or usage billing Model/tool usage and cloud Orbs are billed separately; workspace and tier restrictions apply. [Check current plans and usage terms ↗](https://ampcode.com/docs/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://ampcode.com/docs/cli) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://ampcode.com/docs) Editor integration [Documented ↗](https://ampcode.com/docs) Desktop app [Documented ↗](https://ampcode.com/docs) Web interface [Documented ↗](https://ampcode.com/docs) Open source Unknown The reviewed sources did not establish this capability for a specific interface. Multiple model providers [Multiple providers ↗](https://ampcode.com/docs/customize/model-routing) Amp documents shared agents and threads across its CLI, editor integration, native apps and web. Model routing, skill sources and MCP configuration still depend on the account and execution environment. Scope: cli, ide, desktop, web Local model connection Unknown The reviewed sources did not establish this capability for a specific interface. MCP [Documented ↗](https://ampcode.com/docs/customize/mcp) Amp documents shared agents and threads across its CLI, editor integration, native apps and web. Model routing, skill sources and MCP configuration still depend on the account and execution environment. Scope: cli, ide, desktop, web Agent skills [Documented ↗](https://ampcode.com/docs/customize/skills) Amp documents shared agents and threads across its CLI, editor integration, native apps and web. Model routing, skill sources and MCP configuration still depend on the account and execution environment. Scope: cli, ide, desktop, web Subagents [Documented ↗](https://ampcode.com/docs/plugin-api) CLI plugins can define specialized agents and expose delegation through a tool. This scope does not establish equivalent controls in every Amp interface. Scope: cli Hooks / extensions [Documented ↗](https://ampcode.com/docs/plugin-api) Documented CLI hooks use TypeScript/JavaScript plugin events for tool calls and agent lifecycle; this is not a claim of compatibility with other products' hook files. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://ampcode.com/docs) Skills installation target: .agents/skills/ · Global: $XDG_CONFIG_HOME/agents/skills/ (default ~/.config/agents/skills/) . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Introduction ↗](https://ampcode.com/docs) official docs · checked 2026-09-21 [CLI setup ↗](https://ampcode.com/docs/cli) official docs · checked 2026-09-21 [Model routing ↗](https://ampcode.com/docs/customize/model-routing) official docs · checked 2026-09-21 [Plugin API ↗](https://ampcode.com/docs/plugin-api) official docs · checked 2026-09-21 [Pricing ↗](https://ampcode.com/docs/pricing) official docs · checked 2026-09-21 [Skills and their availability ↗](https://ampcode.com/docs/customize/skills) official docs · checked 2026-09-21 [MCP in local and cloud executors ↗](https://ampcode.com/docs/customize/mcp) official docs · checked 2026-09-21 [Plugins and event behavior ↗](https://ampcode.com/docs/customize/plugins) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Antigravity](#antigravity) Coding editor · Google Google development environment with agent review artifacts, browser work and an editor interface. | IDE | Multiple providers Local connection unknown | Licence not established Plan quotas and AI credits | Compare Antigravity Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Antigravity Product mark · Profile reviewed 2026-09-21 Why consider Antigravity Work with code, browser actions and review artifacts in one environment. Use documented skills, subagents and hooks to customize agent work. What to weigh up Model availability changes by plan and enterprise policy. Local-weight execution was not established for the IDE in this review. Models & access The IDE offers a plan-dependent hosted model selector including Google, Anthropic and open-weight models. An open-weight model in the selector does not establish on-device inference. Plan quotas and AI credits Free and paid Google AI plans have different model availability and quotas. [Check current plans and usage terms ↗](https://antigravity.google/docs/plans) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://antigravity.google/download) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration [Documented ↗](https://antigravity.google/docs/getting-started) Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers [Multiple providers ↗](https://antigravity.google/docs/models) Scope: ide Local model connection Unknown MCP [Documented ↗](https://antigravity.google/docs/getting-started) Scope: ide Agent skills [Documented ↗](https://antigravity.google/docs/getting-started) Scope: ide Subagents [Documented ↗](https://antigravity.google/docs/getting-started) Scope: ide Hooks / extensions [Documented ↗](https://antigravity.google/docs/getting-started) Scope: ide Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://antigravity.google/docs/getting-started) Skills installation target: .agents/skills/ · Global: ~/.gemini/antigravity/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Antigravity getting started ↗](https://antigravity.google/docs/getting-started) official docs · checked 2026-09-21 [Antigravity models ↗](https://antigravity.google/docs/models) official docs · checked 2026-09-21 [Plans and AI credits ↗](https://antigravity.google/docs/plans) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Antigravity CLI](#antigravity-cli) Coding agent · Google Google’s terminal interface to the Antigravity agent, with shared settings, headless execution and IDE handoff. Also known as agy | CLI | Multiple providers Local connection unknown | Licence not established Included quota; paid plans and overages | Compare Antigravity CLI Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Antigravity CLI Product mark · Profile reviewed 2026-09-21 Why consider Antigravity CLI Use Antigravity from SSH, tmux or a terminal workflow. Carry settings and conversations into the Antigravity visual environment. What to weigh up Model availability changes by plan and enterprise policy. The CLI is a separate surface from the Antigravity IDE; one is not the installer for the other. Models & access Antigravity offers a plan-dependent catalog spanning Gemini and selected third-party models. CLI also documents direct Gemini API-key authentication; arbitrary local-model support was not established. Included quota; paid plans and overages Google-account access includes a baseline quota. Google AI paid plans increase allowances, and eligible subscribers can purchase AI credits for optional usage beyond that quota. Model access and limits depend on the plan. Direct Gemini API-key authentication is a separate CLI access path. [Check current plans and usage terms ↗](https://antigravity.google/docs/plans) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://antigravity.google/docs/cli/install) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://antigravity.google/docs/cli/overview) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers [Multiple providers ↗](https://antigravity.google/docs/cli/install) Local model connection Unknown MCP [Documented ↗](https://antigravity.google/docs/cli/overview) Agent skills [Documented ↗](https://antigravity.google/docs/cli/overview) Subagents [Documented ↗](https://antigravity.google/docs/cli/overview) Hooks / extensions [Documented ↗](https://antigravity.google/docs/cli/overview) Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://antigravity.google/docs/cli/overview) Skills installation target: .agents/skills/ · Global: ~/.gemini/antigravity-cli/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [CLI overview ↗](https://antigravity.google/docs/cli/overview) official docs · checked 2026-09-21 [Installation and authentication ↗](https://antigravity.google/docs/cli/install) official docs · checked 2026-09-21 [Models ↗](https://antigravity.google/docs/models/) official docs · checked 2026-09-21 [Antigravity plans, quotas and credit overages ↗](https://antigravity.google/docs/plans) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [AnythingLLM](#anythingllm) Model client · Mintplex Labs Document and agent workspace with local or cloud models, desktop and self-hosted deployment. | desktop web | Multiple providers Local connection documented | [MIT ↗](https://github.com/Mintplex-Labs/anything-llm/blob/master/LICENSE) Open-source self-hosting; hosted option | Compare AnythingLLM Add |
| Reasons, limits, setup & sources 1 sources · 2026-09-21 AnythingLLM Product mark · Profile reviewed 2026-09-21 Why consider AnythingLLM Build document-grounded workspaces with citations and agent tools. Choose a desktop installation or a self-hosted multi-user instance. What to weigh up Multi-user permissions and embeddable widgets are documented for Docker, not desktop parity. Local privacy depends on the selected model, embedding and other providers. Models & access Connects local runtimes including Ollama and LM Studio or cloud providers. Supports document ingestion, agent workflows and configurable model routing. Open-source self-hosting; hosted option Hosting and chosen provider usage are separate costs. [Check current plans and usage terms ↗](https://github.com/Mintplex-Labs/anything-llm) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://anythingllm.com/download) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://github.com/Mintplex-Labs/anything-llm) Web interface [Documented ↗](https://github.com/Mintplex-Labs/anything-llm) Open source [Open source ↗](https://github.com/Mintplex-Labs/anything-llm) Scope: desktop, web Multiple model providers [Multiple providers ↗](https://github.com/Mintplex-Labs/anything-llm) Scope: desktop, web Local model connection [Documented ↗](https://github.com/Mintplex-Labs/anything-llm) Desktop/server connects to a configured local model runtime; a remote web server does not imply inference on the browser device. Scope: desktop, web MCP [Documented ↗](https://github.com/Mintplex-Labs/anything-llm) Scope: desktop, web Agent skills Unknown Its agent-tool skills were not equated with portable SKILL.md support. Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://github.com/Mintplex-Labs/anything-llm) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [AnythingLLM official repository ↗](https://github.com/Mintplex-Labs/anything-llm) official repository · checked 2026-09-21 | | | | |
| [AstrBot](#astrbot) AI assistant · AstrBot contributors Self-hosted assistant connected to messaging platforms, with web chat and an agent sandbox. | web desktop messaging | Multiple providers Local connection documented | [AGPL-3.0 ↗](https://github.com/AstrBotDevs/AstrBot/blob/master/LICENSE) Self-hosted software | Compare AstrBot Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 AstrBot Product mark · Profile reviewed 2026-09-21 Why consider AstrBot Chat-platform assistants using your own model service. Web chat with plugins and isolated code execution. What to weigh up Requires deployment and configuration of model and messaging connections. Desktop deployment is explicitly discouraged for server use. Models & access Official provider list includes hosted APIs, OpenAI-compatible services, Ollama and LM Studio. Self-hosted software Model-service and hosting costs depend on your deployment. [Check current plans and usage terms ↗](https://github.com/AstrBotDevs/AstrBot) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/AstrBotDevs/AstrBot) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://github.com/AstrBotDevs/AstrBot) Web interface [Documented ↗](https://github.com/AstrBotDevs/AstrBot) Open source [Open source ↗](https://github.com/AstrBotDevs/AstrBot/blob/master/LICENSE) Scope: web, messaging Multiple model providers [Multiple providers ↗](https://github.com/AstrBotDevs/AstrBot) Scope: web, messaging Local model connection [Documented ↗](https://github.com/AstrBotDevs/AstrBot) Scope: web, messaging MCP [Documented ↗](https://github.com/AstrBotDevs/AstrBot) Scope: web, messaging Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Scope: web, messaging Subagents Unknown Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/AstrBotDevs/AstrBot) Skills installation target: data/skills/ · Global: ~/.astrbot/data/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Astrbot Readme ↗](https://github.com/AstrBotDevs/AstrBot) official repository · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [AGPL-3.0 license ↗](https://github.com/AstrBotDevs/AstrBot/blob/master/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Augment](#augment) Framework / platform · Augment Code Codebase-aware assistants and development automation through IDE extensions, Auggie CLI and Cosmos. | IDE CLI web | Multiple providers Local connection unknown | Licence not established Tokens, service fee and cloud compute | Compare Augment Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Augment Product mark · Profile reviewed 2026-09-21 Why consider Augment Use an indexed codebase context service inside VS Code or JetBrains. Run development automation from triggers and schedules through Cosmos. What to weigh up Token, service-fee and cloud-compute charges can all contribute to cost. Subagents and hooks are documented for CLI/Cosmos; do not assume full extension parity. Models & access Hosted model selection spans multiple providers; the last selection or organization default is used. Local model support was not confirmed. Tokens, service fee and cloud compute Current Cosmos/Auggie pricing includes model usage, a service fee and applicable compute; see product-specific terms. [Check current plans and usage terms ↗](https://docs.augmentcode.com/models/token-based-pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.augmentcode.com/setup-augment/install-visual-studio-code) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://docs.augmentcode.com/introduction) Editor integration [Documented ↗](https://docs.augmentcode.com/introduction) Desktop app Unknown Web interface [Documented ↗](https://docs.augmentcode.com/introduction) Open source Unknown Multiple model providers [Multiple providers ↗](https://docs.augmentcode.com/models/available-models) Model catalogue and token billing verified for Auggie CLI/Cosmos; no arbitrary local inference verified. Scope: cli, web Local model connection Unknown MCP [Documented ↗](https://docs.augmentcode.com/llms.txt) Scope: ide, cli, web Agent skills [Documented ↗](https://docs.augmentcode.com/llms.txt) Scope: ide, cli, web Subagents [Documented ↗](https://docs.augmentcode.com/llms.txt) Scope: cli, web Hooks / extensions [Documented ↗](https://docs.augmentcode.com/llms.txt) CLI execution hooks; Cosmos event webhooks are a different capability. Scope: cli Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://docs.augmentcode.com/introduction) Skills installation target: .augment/skills/ · Global: ~/.augment/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Augment introduction ↗](https://docs.augmentcode.com/introduction) official docs · checked 2026-09-21 [Augment official documentation index ↗](https://docs.augmentcode.com/llms.txt) official docs · checked 2026-09-21 [Available models ↗](https://docs.augmentcode.com/models/available-models) official docs · checked 2026-09-21 [Token-based pricing ↗](https://docs.augmentcode.com/models/token-based-pricing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Autohand Code](#autohand-code) Coding agent · Autohand AI Coding workspace with terminal, editor and web surfaces, configurable providers and extensible agent workflows. | CLI IDE web | Multiple providers Local connection documented | [License conflict: Apache-2.0 / commercial terms ↗](https://github.com/autohandai/code-cli/blob/main/COMMERCIAL.md) Provider costs; commercial terms vary | Compare Autohand Code Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Autohand Code Product mark · Profile reviewed 2026-09-21 Why consider Autohand Code Use local inference options or hosted providers from a customizable terminal agent. Extend workflows through skills, hooks, MCP and agent definitions. What to weigh up LICENSE is Apache-2.0, but README and COMMERCIAL.md impose a revenue-based commercial restriction. The licensing conflict needs clarification before calling the runtime unrestricted open source. Models & access Documents hosted providers, Ollama, llama.cpp and MLX, with provider switching. Provider costs; commercial terms vary Hosted model providers bill through your configured accounts; local backends use your own compute. The repository describes no-charge use for some users and a commercial license requirement for certain organizations, but its licensing notices conflict. Check both notices before assuming organization-wide client use is free. [Check current plans and usage terms ↗](https://github.com/autohandai/code-cli/blob/main/COMMERCIAL.md) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/autohandai/code-cli) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/autohandai/code-cli) Editor integration [Documented ↗](https://github.com/autohandai/code-cli) Desktop app Unknown Web interface [Documented ↗](https://github.com/autohandai/code-cli) Open source Unknown The Apache licence file conflicts with revenue-based restrictions in COMMERCIAL.md; unrestricted open-source status remains unresolved for every surface. Multiple model providers [Multiple providers ↗](https://github.com/autohandai/code-cli) The model-access evidence in this profile covers the terminal client. Other listed product surfaces must be checked separately. Documents hosted providers, Ollama, llama.cpp and MLX, with provider switching. Scope: cli Local model connection [Documented ↗](https://github.com/autohandai/code-cli) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli MCP [Documented ↗](https://github.com/autohandai/code-cli) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Agent skills [Documented ↗](https://github.com/autohandai/code-cli) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents [Documented ↗](https://github.com/autohandai/code-cli) The CLI repository documents spawn_subagent for focused delegated work. The separate web and editor integrations are not independently verified by this source. Scope: cli Hooks / extensions [Documented ↗](https://github.com/autohandai/code-cli) The CLI documents extension hooks and lifecycle integrations. These should not be confused with its separate Git-hook command or assumed available in every client. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/autohandai/code-cli) Skills installation target: .autohand/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Autohand CLI README ↗](https://github.com/autohandai/code-cli) official repository · checked 2026-09-21 [Apache license file ↗](https://github.com/autohandai/code-cli/blob/main/LICENSE) official repository · checked 2026-09-21 [Additional commercial terms ↗](https://github.com/autohandai/code-cli/blob/main/COMMERCIAL.md) official repository · checked 2026-09-21 [Product surfaces ↗](https://docs.autohand.ai/) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Bolt](#bolt) Framework / platform · StackBlitz Browser-based app builder with code editing, previews, integrations and hosted publishing. | web | Not established Local connection unknown | Licence not established Token-based plans and hosted services | Compare Bolt Add |
| Reasons, limits, setup & sources 1 sources · 2026-09-21 Bolt Product mark · Profile reviewed 2026-09-21 Why consider Bolt Turn a design or prompt into an editable web application in the browser. Use built-in hosting/database workflows or import an existing GitHub project. What to weigh up Focuses on supported web stacks; it is not a general local shell agent. Token use, hosting and production services need separate cost review. Models & access Hosted agent selection is documented; no arbitrary local inference endpoint verified. Token-based plans and hosted services Build usage is token-based; deployed application services can have separate plan constraints. [Check current plans and usage terms ↗](https://support.bolt.new/get-started/intro-bolt) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://support.bolt.new/get-started/intro-bolt) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app Unknown Web interface [Documented ↗](https://support.bolt.new/get-started/intro-bolt) Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP [Documented ↗](https://support.bolt.new/get-started/intro-bolt) Scope: web Agent skills [Documented ↗](https://support.bolt.new/get-started/intro-bolt) Scope: web Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://support.bolt.new/get-started/intro-bolt) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Bolt introduction ↗](https://support.bolt.new/get-started/intro-bolt) official docs · checked 2026-09-21 | | | | |
| [ChatGPT](#chatgpt) AI assistant · OpenAI OpenAI assistant for conversation, research and producing files, with Work tasks on the web and desktop and separate Codex developer views. | desktop web mobile | Not established Local connection unknown | Licence not established Free access and paid plans; usage limits apply | Compare ChatGPT Add |
| Reasons, limits, setup & sources 7 sources · 2026-09-21 ChatGPT Product mark · Profile reviewed 2026-09-21 Why consider ChatGPT Research sources, analyze uploaded material and produce documents or spreadsheets in a managed workspace. Use desktop file/app access and inspect finished artifacts; use web Work for cloud tasks that can continue after closing the app. What to weigh up Features vary by plan, rollout, region and workspace policy; Work and Codex share usage. Local task execution and local MCP access do not establish local model inference or an offline assistant. Models & access Chat and Work expose plan-dependent OpenAI models. Desktop Codex views also share Codex configuration, but their provider options are not attributed to ordinary Chat or voice. Free access and paid plans; usage limits apply Work and Codex share usage and credits. API access is a separate billing path. [Check current plans and usage terms ↗](https://learn.chatgpt.com/docs/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://learn.chatgpt.com/docs/app) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://learn.chatgpt.com/docs/app.md) Web interface [Documented ↗](https://learn.chatgpt.com/docs/app.md) Open source Unknown Multiple model providers Unknown Local model connection Unknown Local files and tool execution are separate from model inference; no offline ChatGPT claim is made. MCP [Documented ↗](https://learn.chatgpt.com/docs/extend/mcp.md) Web uses remote MCP tools from plugins. Desktop Codex-host configuration supports local stdio and Streamable HTTP. Web does not read local config files. Scope: desktop, web Agent skills [Documented ↗](https://learn.chatgpt.com/docs/skills-and-plugins.md) ChatGPT skills/plugins are available where enabled; workspace administrators and rollout control access. Scope: desktop, web Subagents [Documented ↗](https://learn.chatgpt.com/docs/use-chatgpt.md) Work tasks can delegate; this is not a claim that every ordinary Chat response uses subagents. Scope: desktop, web Hooks / extensions [Documented ↗](https://learn.chatgpt.com/docs/skills-and-plugins.md) Hooks require a Codex-runtime execution environment, including Work; installing a web plugin does not deploy its local scripts. Scope: desktop, web Availability: Current official product documentation verified. [Source ↗](https://chatgpt.com) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [ChatGPT desktop app ↗](https://learn.chatgpt.com/docs/app.md) official docs · checked 2026-09-21 [Use ChatGPT and Work ↗](https://learn.chatgpt.com/docs/use-chatgpt.md) official docs · checked 2026-09-21 [ChatGPT Work and Codex model scope ↗](https://learn.chatgpt.com/docs/models.md) official docs · checked 2026-09-21 [ChatGPT web and desktop MCP scope ↗](https://learn.chatgpt.com/docs/extend/mcp.md) official docs · checked 2026-09-21 [Skills and plugins in ChatGPT ↗](https://learn.chatgpt.com/docs/skills-and-plugins.md) official docs · checked 2026-09-21 [ChatGPT Work pricing and shared usage ↗](https://learn.chatgpt.com/docs/pricing.md) official docs · checked 2026-09-21 [OpenAI plugin execution and surface boundaries ↗](https://developers.openai.com/plugins/concepts/plugins) official docs · checked 2026-09-21 | | | | |
| [ChatWise](#chatwise) Model client · ChatWise Desktop multi-provider chat client with local history, model endpoint configuration, MCP tools and a separately documented Agent preview. | desktop | Multiple providers Local connection documented | Licence not established Personal free tier; paid Pro license plus model usage | Compare ChatWise Add |
| Reasons, limits, setup & sources 6 sources · 2026-09-21 ChatWise Product mark · Profile reviewed 2026-09-21 Why consider ChatWise Keep a desktop chat workspace while switching among cloud APIs and separately served local models. Use Agent preview with project files, portable SKILL.md packages and delegated exploration. What to weigh up Agent preview explicitly disables MCP and Artifacts and currently has no sandbox. Normal chat MCP lacks OAuth and server-initiated Streamable HTTP messages; docs also warn local Ollama tool use is not optimized. Models & access Bring provider keys or connect compatible endpoints, including Ollama and LM Studio. Requests go to the chosen model provider. Personal free tier; paid Pro license plus model usage Free access is limited to personal non-commercial use. Pro is a one-time license with a bounded feature-update period; model access is separate. [Check current plans and usage terms ↗](https://chatwise.app/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.chatwise.app/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://chatwise.app) Web interface Unknown Open source Unknown Multiple model providers [Multiple providers ↗](https://chatwise.app) Scope: desktop Local model connection [Documented ↗](https://docs.chatwise.app/ollama) ChatWise connects to a separately installed local runtime. Hosted provider calls and enabled online tools still leave the device. Scope: desktop MCP [Documented ↗](https://docs.chatwise.app/tools) Normal chat supports stdio, SSE and limited Streamable HTTP; no OAuth. MCP is disabled in the documented Agent preview. Scope: desktop Agent skills [Documented ↗](https://docs.chatwise.app/agent) Portable SKILL.md support is in Agent preview, which disables MCP and has no sandbox. Scope: desktop Subagents [Documented ↗](https://docs.chatwise.app/agent) Agent preview supports general-purpose and exploration subagents; general-purpose subagents do not get Skills. Scope: desktop Hooks / extensions Unknown Combinations to check Not available together: Portable Skills are documented in Agent preview, which explicitly disables MCP. Normal-chat MCP and Agent Skills cannot be combined in that documented mode. [Source ↗](https://docs.chatwise.app/agent) Not available together: Subagents are documented in Agent preview, which explicitly disables MCP. The independently supported features do not form one compatible workflow. [Source ↗](https://docs.chatwise.app/agent) Availability: Current official product documentation verified. [Source ↗](https://chatwise.app) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [ChatWise product and data routing ↗](https://chatwise.app) official docs · checked 2026-09-21 [ChatWise compatible provider setup ↗](https://docs.chatwise.app/custom-provider) official docs · checked 2026-09-21 [ChatWise local Ollama connection ↗](https://docs.chatwise.app/ollama) official docs · checked 2026-09-21 [ChatWise MCP transports and limits ↗](https://docs.chatwise.app/tools) official docs · checked 2026-09-21 [ChatWise Agent skills/subagents and exclusions ↗](https://docs.chatwise.app/agent) official docs · checked 2026-09-21 [ChatWise license purchase and free-tier scope ↗](https://chatwise.app/pricing) official docs · checked 2026-09-21 | | | | |
| [Cherry Studio](#cherry-studio) Model client · Cherry Studio Desktop client combining provider choice, local endpoints, document tools and parallel model conversations. | desktop | Multiple providers Local connection documented | [AGPL-3.0 ↗](https://github.com/CherryHQ/cherry-studio/blob/main/LICENSE) Community client; commercial edition | Compare Cherry Studio Add |
| Reasons, limits, setup & sources 1 sources · 2026-09-21 Cherry Studio Product mark · Profile reviewed 2026-09-21 Why consider Cherry Studio Compare several model responses inside one desktop conversation workflow. Use custom assistants and documents across local and cloud providers. What to weigh up Community and commercial editions have different terms. Cloud calls and integrations have different privacy boundaries from local inference. Models & access Cloud providers plus local Ollama and LM Studio connections; multi-model simultaneous conversations are documented. Community client; commercial edition The community edition is AGPL; provider use and commercial-license terms are separate. [Check current plans and usage terms ↗](https://github.com/CherryHQ/cherry-studio) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/CherryHQ/cherry-studio/releases) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://raw.githubusercontent.com/CherryHQ/cherry-studio/main/README.md) Web interface Unknown Open source [Open source ↗](https://raw.githubusercontent.com/CherryHQ/cherry-studio/main/README.md) Scope: desktop Multiple model providers [Multiple providers ↗](https://raw.githubusercontent.com/CherryHQ/cherry-studio/main/README.md) Scope: desktop Local model connection [Documented ↗](https://raw.githubusercontent.com/CherryHQ/cherry-studio/main/README.md) Scope: desktop MCP [Documented ↗](https://raw.githubusercontent.com/CherryHQ/cherry-studio/main/README.md) Scope: desktop Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://raw.githubusercontent.com/CherryHQ/cherry-studio/main/README.md) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Cherry Studio official README ↗](https://raw.githubusercontent.com/CherryHQ/cherry-studio/main/README.md) official repository · checked 2026-09-21 | | | | |
| [Claude](#claude) AI assistant · Anthropic Anthropic assistant across web, desktop and mobile for conversation, connected research and producing files; the former Cowork experience is being merged into chat. Also known as Claude Desktop · Claude Cowork | desktop web mobile | Not established Local connection unknown | Licence not established Free access and paid subscriptions | Compare Claude Add |
| Reasons, limits, setup & sources 6 sources · 2026-09-21 Claude Product mark · Profile reviewed 2026-09-21 Why consider Claude Connect work services and create editable reports, spreadsheets and presentations. Use Desktop for access to permitted local folders, local MCP tools and supported computer workflows. What to weigh up Unified chat/Cowork is rolling out by plan and account; older accounts still show separate modes. Remote MCP originates from Anthropic cloud and needs a publicly reachable server; local desktop extensions have a different connection path. Models & access Official consumer plans offer Claude models. Optional externally configured desktop routing is not verified as a general feature here; local files and connectors do not imply local model weights. Free access and paid subscriptions Usage, models and advanced work features depend on plan; enterprise terms differ. [Check current plans and usage terms ↗](https://claude.com/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://claude.com/download) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://claude.com/download) Web interface [Documented ↗](https://claude.com/download) Open source Unknown Multiple model providers Unknown Local model connection Unknown Access to files on your computer does not mean Claude model inference runs there. MCP [Documented ↗](https://support.claude.com/en/articles/10949351-getting-started-with-local-mcp-servers-on-claude-desktop) Desktop supports local MCP extensions. Remote connectors on web, desktop and mobile connect from Anthropic cloud, not your local network. Scope: desktop, web, mobile Agent skills [Documented ↗](https://support.claude.com/en/articles/13837440-use-plugins-in-claude) Plugin skills are documented for web chat, Desktop chat and Cowork; plan and organization policies apply. Scope: desktop, web Subagents [Documented ↗](https://support.claude.com/en/articles/13837440-use-plugins-in-claude) Explicitly documented for Cowork, with unified chat rollout still in progress; not guaranteed in ordinary chat. Scope: desktop Hooks / extensions [Documented ↗](https://support.claude.com/en/articles/13837440-use-plugins-in-claude) Plugin hooks are documented for Cowork, not ordinary chat. Scope: desktop Availability: Current official product documentation verified. [Source ↗](https://claude.ai) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Claude applications and platform limits ↗](https://claude.com/download) official docs · checked 2026-09-21 [Claude and Cowork unified experience rollout ↗](https://support.claude.com/en/articles/16761823-claude-cowork-and-chat-are-one-claude) official docs · checked 2026-09-21 [Plugins, skills and Cowork-only capabilities ↗](https://support.claude.com/en/articles/13837440-use-plugins-in-claude) official docs · checked 2026-09-21 [Claude local desktop MCP ↗](https://support.claude.com/en/articles/10949351-getting-started-with-local-mcp-servers-on-claude-desktop) official docs · checked 2026-09-21 [Claude remote MCP network requirements ↗](https://support.claude.com/en/articles/11175166-get-started-with-custom-connectors-using-remote-mcp) official docs · checked 2026-09-21 [Claude plans and model access ↗](https://claude.com/pricing) official docs · checked 2026-09-21 | | | | |
| [Claude Code](#claude-code) Coding agent · Anthropic Anthropic’s coding agent for terminal, editor, desktop and browser workflows, with extensible skills and lifecycle controls. | CLI IDE desktop web | Provider ecosystem No local model connection | [proprietary ↗](https://github.com/anthropics/claude-code/blob/main/LICENSE.md) Claude plan or metered API | Compare Claude Code Add |
| Reasons, limits, setup & sources 6 sources · 2026-09-21 Claude Code Product mark · Profile reviewed 2026-09-21 Why consider Claude Code Use Claude with project instructions, skills, MCP and specialized agents. Apply lifecycle hooks and organization-managed settings to coding workflows. What to weigh up Official model choice is Claude; cloud hosting options are not unrestricted model choice. Included usage and defaults depend on the account and provider. Models & access Official model configuration targets Claude through Anthropic or supported cloud deployments. Changing ANTHROPIC_BASE_URL changes the endpoint, not the model family; runtime defaults depend on account/provider. Claude plan or metered API The sign-in method determines billing. Claude subscriptions and organization seats include usage allowances, with optional usage credits on eligible plans. Anthropic Console and supported cloud-provider access are billed by token to the corresponding account; a Claude subscription does not fund separate API-key usage. [Check current plans and usage terms ↗](https://code.claude.com/docs/en/costs) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://code.claude.com/docs/en/overview) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://code.claude.com/docs/en/overview) Editor integration [Documented ↗](https://code.claude.com/docs/en/overview) Desktop app [Documented ↗](https://code.claude.com/docs/en/overview) Web interface [Documented ↗](https://code.claude.com/docs/en/overview) Open source [Not open source ↗](https://github.com/anthropics/claude-code/blob/main/LICENSE.md) The reviewed Claude Code CLI licence is proprietary; a public issue/release repository does not make its runtime open source. Scope: cli Multiple model providers [One provider ecosystem ↗](https://code.claude.com/docs/en/model-config) The official CLI supports the Claude model family through supported hosting providers; hosting choice is not arbitrary model-family choice. Scope: cli Local model connection [No supported local connection ↗](https://code.claude.com/docs/en/model-config) The official Claude CLI model configuration does not document local open-weight inference. This does not adjudicate unofficial compatibility proxies. Scope: cli MCP [Documented ↗](https://code.claude.com/docs/en/overview) Documented for the CLI and Claude Desktop’s Code tab. Desktop connectors, project configuration and cloud/SSH sessions have different setup paths; this does not describe the Chat or Cowork tabs. Scope: cli, desktop Agent skills [Documented ↗](https://code.claude.com/docs/en/overview) Documented for the CLI and Claude Desktop’s Code tab. Desktop connectors, project configuration and cloud/SSH sessions have different setup paths; this does not describe the Chat or Cowork tabs. Scope: cli, desktop Subagents [Documented ↗](https://code.claude.com/docs/en/desktop) Subagents are documented in the CLI and desktop Code sessions; the desktop tasks pane exposes their activity. This does not establish Claude Chat or Cowork behavior. Scope: cli, desktop Hooks / extensions [Documented ↗](https://code.claude.com/docs/en/desktop) The desktop Code documentation explicitly shares hooks and skills settings with the CLI for local/SSH workflows. Do not infer identical cloud-session configuration. Scope: cli, desktop Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://code.claude.com/docs/en/overview) Skills installation target: .claude/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Claude Code overview ↗](https://code.claude.com/docs/en/overview) official docs · checked 2026-09-21 [Model configuration ↗](https://code.claude.com/docs/en/model-config) official docs · checked 2026-09-21 [Official license ↗](https://github.com/anthropics/claude-code/blob/main/LICENSE.md) official repository · checked 2026-09-21 [Claude Desktop Code tab capabilities and CLI comparison ↗](https://code.claude.com/docs/en/desktop) official docs · checked 2026-09-21 [Claude Code costs and subscription versus API billing ↗](https://code.claude.com/docs/en/costs) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Cline](#cline) Coding agent · Cline Model-flexible coding agent available through editors, terminal, desktop and a task board. | IDE CLI desktop web | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/cline/cline/blob/main/LICENSE) BYOK, usage billing or ClinePass | Compare Cline Add |
| Reasons, limits, setup & sources 9 sources · 2026-09-21 Cline Product mark · Profile reviewed 2026-09-21 Why consider Cline Keep an existing editor while choosing model providers and local inference. Use approval controls, MCP and reusable skills across documented Cline surfaces. What to weigh up Inference cost depends on provider, model and work performed. Desktop availability differs from CLI: official docs list macOS and Windows beta; Linux Desktop is not confirmed. Models & access BYOK cloud providers, OpenAI-compatible endpoints and local runtimes including Ollama and LM Studio. Cline-hosted usage and ClinePass are alternative billing paths. BYOK, usage billing or ClinePass The open-source client does not make paid inference free. Promotional free models can change. [Check current plans and usage terms ↗](https://docs.cline.bot/cline-overview) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.cline.bot/cline-overview) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://docs.cline.bot/cline-overview) Editor integration [Documented ↗](https://docs.cline.bot/cline-overview) Desktop app [Documented ↗](https://docs.cline.bot/cline-overview) Web interface [Documented ↗](https://docs.cline.bot/cline-overview) Open source [Open source ↗](https://github.com/cline/cline) CLI, Desktop and VS Code extension are in the open-source repository. README explicitly excludes the JetBrains plugin from open source. Scope: cli, ide, desktop Multiple model providers [Multiple providers ↗](https://docs.cline.bot/cline-overview) Scope: cli, ide, desktop Local model connection [Documented ↗](https://docs.cline.bot/llms.txt) Local runtime setup verified for IDE and Desktop; CLI-specific local setup was not fully verified. Scope: ide, desktop MCP [Documented ↗](https://docs.cline.bot/usage/cli-overview.md) CLI has its own MCP management command; Desktop Marketplace and IDE MCP documented. Web task-board parity not established. Scope: cli, ide, desktop Agent skills [Documented ↗](https://docs.cline.bot/customization/skills.md) IDE Skills panel and Desktop Marketplace verified; portable Skills parity across CLI/web is not assumed. Scope: ide, desktop Subagents [Documented ↗](https://docs.cline.bot/llms.txt) Scope: ide Hooks / extensions [Documented ↗](https://docs.cline.bot/llms.txt) Scope: ide Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://docs.cline.bot/cline-overview) Skills installation target: .agents/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Cline overview ↗](https://docs.cline.bot/cline-overview) official docs · checked 2026-09-21 [Cline documentation index ↗](https://docs.cline.bot/llms.txt) official docs · checked 2026-09-21 [Cline Desktop ↗](https://docs.cline.bot/usage/cline-desktop) official docs · checked 2026-09-21 [Cline source repository ↗](https://github.com/cline/cline) official repository · checked 2026-09-21 [Cline current Apache license ↗](https://raw.githubusercontent.com/cline/cline/main/LICENSE) official repository · checked 2026-09-21 [Cline CLI overview ↗](https://docs.cline.bot/usage/cli-overview.md) official docs · checked 2026-09-21 [Cline local runtime setup ↗](https://docs.cline.bot/running-models-locally/overview.md) official docs · checked 2026-09-21 [Cline IDE Skills panel ↗](https://docs.cline.bot/customization/skills.md) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [CodeArts Agent](#codearts-agent) Coding editor · Huawei Cloud Huawei Cloud coding IDE with agent and specification-driven workflows. | IDE | Not established Local connection unknown | Licence not established Plan seats and token allowances | Compare CodeArts Agent Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 CodeArts Agent Huawei Cloud mark · Profile reviewed 2026-09-21 Why consider CodeArts Agent Use requirements, implementation design and task planning as explicit coding stages. Assess centralized skills and model management for a Huawei Cloud organization. What to weigh up Features, token allowances and governance depend on plan. Regional download and plan pages differ; verify the intended service region. Models & access A managed model selector is documented. Enterprise plans list custom model management, but on-device inference is not verified. Plan seats and token allowances Free and paid editions have different allowances and enterprise capabilities. [Check current plans and usage terms ↗](https://www.huaweicloud.com/intl/en-us/product/codearts/ai.html) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://codearts.huaweicloud.com/download.html) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration [Documented ↗](https://support.huaweicloud.com/intl/en-us/usermanual-codeartsagent/codeartsagent_ug_0005.html) Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP [Documented ↗](https://www.huaweicloud.com/intl/en-us/product/codearts/ai.html) Scope: ide Agent skills [Documented ↗](https://www.huaweicloud.com/intl/en-us/product/codearts/ai.html) Scope: ide Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://support.huaweicloud.com/intl/en-us/usermanual-codeartsagent/codeartsagent_ug_0005.html) Skills installation target: .codeartsdoer/skills/ · Global: ~/.codeartsdoer/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [CodeArts Agent chats ↗](https://support.huaweicloud.com/intl/en-us/usermanual-codeartsagent/codeartsagent_ug_0005.html) official docs · checked 2026-09-21 [CodeArts Agent product and plans ↗](https://www.huaweicloud.com/intl/en-us/product/codearts/ai.html) official docs · checked 2026-09-21 [CodeArts download centre ↗](https://codearts.huaweicloud.com/download.html) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [CodeBuddy](#codebuddy) Coding editor · Tencent Tencent development tool with IDE workflows, model configuration and external integrations. | IDE CLI | Multiple providers Local connection unknown | Licence not established Free allowance and subscription credits | Compare CodeBuddy Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 CodeBuddy Product mark · Profile reviewed 2026-09-21 Why consider CodeBuddy Move from design/plan to implementation with IDE integrations. Configure external services through MCP and custom API models. What to weigh up Promotional credits and feature allowances can expire. CodeBuddy, its CLI and WorkBuddy have distinct surfaces; do not mix their pricing tables. Models & access IDE supports custom model API integration; the fetched overview does not establish on-device weight execution. Free allowance and subscription credits Pricing lists plan-specific credit and model allowances, including time-limited promotional benefits. [Check current plans and usage terms ↗](https://www.codebuddy.ai/pricing/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://www.codebuddy.ai/docs/ide/User-guide/Overview) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://www.codebuddy.ai/docs/ide/User-guide/Overview) Editor integration [Documented ↗](https://www.codebuddy.ai/docs/ide/User-guide/Overview) Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers [Multiple providers ↗](https://www.codebuddy.ai/docs/ide/User-guide/Overview) Custom models documented in CodeBuddy IDE; CLI parity not inferred. Scope: ide Local model connection Unknown MCP [Documented ↗](https://www.codebuddy.ai/docs/ide/User-guide/Overview) Scope: ide Agent skills [Documented ↗](https://www.codebuddy.ai/docs/ide/User-guide/Overview) IDE overview verified; CLI parity not inferred. Scope: ide Subagents [Documented ↗](https://www.codebuddy.ai/docs/ide/Features/Plan-Mode) Scope: ide Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://www.codebuddy.ai/docs/ide/User-guide/Overview) Skills installation target: .codebuddy/skills/ · Global: ~/.codebuddy/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [CodeBuddy IDE overview ↗](https://www.codebuddy.ai/docs/ide/User-guide/Overview) official docs · checked 2026-09-21 [CodeBuddy Plan Mode ↗](https://www.codebuddy.ai/docs/ide/Features/Plan-Mode) official docs · checked 2026-09-21 [CodeBuddy CLI IDE integrations ↗](https://www.codebuddy.ai/docs/cli/ide-integrations) official docs · checked 2026-09-21 [CodeBuddy pricing ↗](https://www.codebuddy.ai/pricing/) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Codemaker](#codemaker) unverified AI assistant · Unknown Skills-registry entry whose exact product identity remains unverified. | Not established | Not established Local connection unknown | Licence not established Not independently verified | Compare Codemaker Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Codemaker Neutral lettermark; product logo not established · Profile reviewed 2026-09-21 Why consider Codemaker No source-verified recommendation is available for this record yet. What to weigh up Do not choose on the basis of this incomplete registry entry. A similarly named CodeMaker AI service documents IDE plugins, but its identity has not been tied to this registry key. Models & access Model access has not been established from the sources reviewed. Not independently verified Check the official service terms before choosing a plan. Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.codemaker.ai/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP Unknown Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: The skills registry key has not been securely mapped to the similarly named official product. [Source ↗](https://docs.codemaker.ai/) Skills installation target: .codemaker/skills/ · Global: ~/.codemaker/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [CodeMaker AI documentation candidate ↗](https://docs.codemaker.ai/) official docs · checked 2026-09-21 [CodeMaker AI VS Code candidate ↗](https://docs.codemaker.ai/documentation/quick-start/getting-started/visual-studio-code) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Codex](#codex) Coding agent · OpenAI OpenAI’s coding agent with an open-source terminal client, editor integration, desktop app and cloud tasks. | CLI IDE desktop web | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/openai/codex/blob/main/LICENSE) ChatGPT allowance or API usage | Compare Codex Add |
| Reasons, limits, setup & sources 6 sources · 2026-09-21 Codex OpenAI mark · Profile reviewed 2026-09-21 Why consider Codex Work across local terminal, editor and cloud coding sessions. Customize the open-source CLI, provider configuration, skills, subagents and trusted hooks. What to weigh up The CLI license does not grant free access to hosted models. Custom endpoints must match the documented API capabilities. Models & access OpenAI is the standard account path. Advanced CLI configuration supports custom model providers and local Ollama-compatible endpoints; compatibility is provider-specific, not proof every feature works with every model. ChatGPT allowance or API usage ChatGPT sign-in uses plan allowances shared with ChatGPT Work; eligible plans can extend usage with credits. API-key mode bills the API account and excludes hosted features such as cloud code review. Configured local or third-party providers have their own inference or compute costs. [Check current plans and usage terms ↗](https://learn.chatgpt.com/docs/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/openai/codex) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/openai/codex) Editor integration [Documented ↗](https://github.com/openai/codex) Desktop app [Documented ↗](https://github.com/openai/codex) Web interface [Documented ↗](https://github.com/openai/codex) Open source [Open source ↗](https://github.com/openai/codex) Apache-2.0 applies to the Codex CLI repository, not automatically to the desktop app or cloud service. Scope: cli Multiple model providers [Multiple providers ↗](https://learn.chatgpt.com/docs/config-file/config-advanced) Verified for the Codex CLI. The desktop app, editor integration and hosted/cloud service have separate scope; this review does not transfer the CLI capability to them. Scope: cli Local model connection [Documented ↗](https://learn.chatgpt.com/docs/config-file/config-advanced) Advanced Codex CLI provider configuration documents local Ollama-compatible endpoints; desktop/cloud parity was not established. Scope: cli MCP [Documented ↗](https://learn.chatgpt.com/docs/config-file/config-advanced) Verified for the Codex CLI. The desktop app, editor integration and hosted/cloud service have separate scope; this review does not transfer the CLI capability to them. Scope: cli Agent skills [Documented ↗](https://learn.chatgpt.com/docs/config-file/config-advanced) Verified for the Codex CLI. The desktop app, editor integration and hosted/cloud service have separate scope; this review does not transfer the CLI capability to them. Scope: cli Subagents [Documented ↗](https://learn.chatgpt.com/docs/agent-configuration/subagents) Official subagent documentation explicitly covers Codex CLI, the IDE extension and the ChatGPT desktop app. Separate hosted ChatGPT Work eligibility is not transferred to this profile's web interface. Scope: cli, ide, desktop Hooks / extensions [Documented ↗](https://learn.chatgpt.com/docs/hooks) The hook documentation establishes local Codex configuration and CLI review/trust controls. Desktop, IDE and hosted configuration parity was not established. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/openai/codex) Skills installation target: .agents/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Codex repository and install guide ↗](https://github.com/openai/codex) official repository · checked 2026-09-21 [Advanced configuration ↗](https://learn.chatgpt.com/docs/config-file/config-advanced) official docs · checked 2026-09-21 [Subagents ↗](https://learn.chatgpt.com/docs/agent-configuration/subagents) official docs · checked 2026-09-21 [Hooks ↗](https://learn.chatgpt.com/docs/hooks) official docs · checked 2026-09-21 [Codex and ChatGPT Work pricing and access paths ↗](https://learn.chatgpt.com/docs/pricing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Command Code](#command-code) Coding agent · Command Code Terminal and desktop coding agent focused on open-model workflows, extensibility and reusable coding preferences. | CLI desktop IDE | Multiple providers Local connection documented | Licence not established Plan credits, BYOK or linked subscriptions | Compare Command Code Add |
| Reasons, limits, setup & sources 6 sources · 2026-09-21 Command Code Product mark · Profile reviewed 2026-09-21 Why consider Command Code Use hosted open models or direct compatible provider/local endpoints. Apply reusable taste preferences, skills and hooks across coding tasks. What to weigh up BYOK is explicitly beta. Vendor claims of best harness performance are not independent comparative evidence. Models & access BYOK beta accepts compatible OpenAI or Anthropic APIs, including local Ollama/vLLM servers. Subscription connections are also documented; routes and usage depend on the provider. Plan credits, BYOK or linked subscriptions Hosted access uses plan credits with rolling limits and optional on-demand top-ups. BYOK, compatible local endpoints and supported linked subscriptions provide separate routes. The documented local-only mode skips Command Code billing reads and routes usage through your chosen provider; provider terms and compute costs still apply. [Check current plans and usage terms ↗](https://commandcode.ai/docs/resources/pricing-limits) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://commandcode.ai/docs) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://commandcode.ai/docs) Editor integration [Documented ↗](https://commandcode.ai/docs) Desktop app [Documented ↗](https://commandcode.ai/docs) Web interface Unknown Open source Unknown The reviewed sources did not establish this capability for a specific interface. Multiple model providers [Multiple providers ↗](https://commandcode.ai/docs/byok) The model-access evidence in this profile covers the terminal client. Other listed product surfaces must be checked separately. BYOK beta accepts compatible OpenAI or Anthropic APIs, including local Ollama/vLLM servers. Subscription connections are also documented; routes and usage depend on the provider. Scope: cli Local model connection [Documented ↗](https://commandcode.ai/docs/byok) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli MCP [Documented ↗](https://commandcode.ai/docs) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Agent skills [Documented ↗](https://commandcode.ai/docs) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents [Documented ↗](https://commandcode.ai/docs/agents) Command Code CLI documents custom agents with separate worker context, tools and model configuration. Desktop and editor parity was not established. Scope: cli Hooks / extensions [Documented ↗](https://commandcode.ai/docs/hooks) The official hooks guide describes commands that run at lifecycle points in a Command Code CLI session. Desktop and editor parity was not established. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://commandcode.ai/docs) Skills installation target: .commandcode/skills/ · Global: ~/.commandcode/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Product documentation ↗](https://commandcode.ai/docs) official docs · checked 2026-09-21 [BYOK providers ↗](https://commandcode.ai/docs/byok) official docs · checked 2026-09-21 [Command Code custom agents ↗](https://commandcode.ai/docs/agents) official docs · checked 2026-09-21 [Command Code lifecycle hooks ↗](https://commandcode.ai/docs/hooks) official docs · checked 2026-09-21 [Command Code subscription credits and usage limits ↗](https://commandcode.ai/docs/resources/pricing-limits) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Continue](#continue) unmaintained Coding agent · Continue Open-source coding assistant whose upstream repository is now read-only and unmaintained. | IDE CLI | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/continuedev/continue/blob/main/LICENSE) Open-source client; inference separate | Compare Continue Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Continue Product mark · Profile reviewed 2026-09-21 Why consider Continue Inspect or reuse an Apache-licensed assistant codebase. Maintain an existing installation with a clear awareness of upstream status. What to weigh up Official repository is read-only and no longer actively maintained. Do not promote as a default new-user recommendation without a maintained successor. Models & access Documentation supports configurable providers and local Ollama/self-hosted models. Upstream is no longer maintained. Open-source client; inference separate Current commercial service availability was not established; do not reuse historical subscription claims. [Check current plans and usage terms ↗](https://github.com/continuedev/continue) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://raw.githubusercontent.com/continuedev/continue/main/README.md) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://raw.githubusercontent.com/continuedev/continue/main/README.md) Editor integration [Documented ↗](https://raw.githubusercontent.com/continuedev/continue/main/README.md) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://raw.githubusercontent.com/continuedev/continue/main/README.md) Open-source code remains available, but maintenance has stopped. Scope: ide, cli Multiple model providers [Multiple providers ↗](https://docs.continue.dev) Scope: ide Local model connection [Documented ↗](https://docs.continue.dev) Legacy IDE configuration; upstream is no longer maintained. Scope: ide MCP [Documented ↗](https://docs.continue.dev) Scope: ide Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Official README states the repository is read-only and no longer maintained after its final release. [Source ↗](https://raw.githubusercontent.com/continuedev/continue/main/README.md) Skills installation target: .continue/skills/ · Global: ~/.continue/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Continue official README ↗](https://raw.githubusercontent.com/continuedev/continue/main/README.md) official repository · checked 2026-09-21 [Continue documentation ↗](https://docs.continue.dev) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Corust Agent](#corust-agent) unverified Coding agent · Corust AI Rust-focused ACP agent distribution; current runtime capabilities and service availability remain unresolved. | IDE | Not established Local connection unknown | [GPL-3.0-or-later ↗](https://github.com/rustingai/corust-agent-release/blob/main/LICENSE) Not established | Compare Corust Agent Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Corust Agent Product mark · Profile reviewed 2026-09-21 Why consider Corust Agent No source-verified recommendation is available for this record yet. What to weigh up Official website returned502 during this pass. Public repository provides release metadata/license but no product README or agent implementation source. Models & access No primary model-selection documentation retrieved. Not established Current account and billing terms are unavailable from the reachable product documentation. [Check current plans and usage terms ↗](https://corust.ai) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/rustingai/corust-agent-release/releases) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration [Documented ↗](https://github.com/rustingai/corust-agent-release) Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP Unknown Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Release artifacts and an ACP extension manifest exist, but current product/service documentation could not be verified. [Source ↗](https://corust.ai) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Corust Agent Readme ↗](https://github.com/rustingai/corust-agent-release) official repository · checked 2026-09-21 [Corust Manifest ↗](https://github.com/rustingai/corust-agent-release/blob/main/extension.toml) official repository · checked 2026-09-21 [Corust Releases ↗](https://github.com/rustingai/corust-agent-release/releases) official repository · checked 2026-09-21 [Corust Agent Acp Discovery ↗](https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json) registry · checked 2026-09-21 [Corust Agent License ↗](https://github.com/rustingai/corust-agent-release/blob/main/LICENSE) official repository · checked 2026-09-21 | | | | |
| [Crow](#crow-cli) Coding agent · Crow contributors ACP-native coding agent with terminal operation, editor integrations and persistent searchable SQL session memory. Also known as crow-cli | CLI IDE | Multiple providers Local connection unknown | [AGPL-3.0-or-later ↗](https://github.com/crow-cli/crow-cli/blob/main/LICENSE.md) Open-source client; provider access separate | Compare Crow Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Crow Product mark · Profile reviewed 2026-09-21 Why consider Crow Resuming coding sessions with searchable local history. Using one agent through terminal and ACP-native editor tools. What to weigh up Web search requires the documented SearXNG/Docker backing service. MCP tool names are not namespaced, so added servers may collide. WSL2 is recommended for Windows setup. Models & access Requires an OpenAI-compatible provider API key; custom endpoints are documented, but local-weight hosting was not explicitly established. Open-source client; provider access separate Provider credentials and backing-service hosting are required as configured. [Check current plans and usage terms ↗](https://github.com/crow-cli/crow-cli) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/crow-cli/crow-cli) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/crow-cli/crow-cli) Editor integration [Documented ↗](https://github.com/crow-cli/crow-cli) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/crow-cli/crow-cli/blob/main/LICENSE.md) Scope: cli, ide Multiple model providers [Multiple providers ↗](https://github.com/crow-cli/crow-cli) Scope: cli, ide Local model connection Unknown MCP [Documented ↗](https://github.com/crow-cli/crow-cli) Scope: cli, ide Agent skills [Documented ↗](https://github.com/crow-cli/crow-cli) Scope: cli, ide Subagents [Documented ↗](https://github.com/crow-cli/crow-cli) The same ACP-native agent supports terminal/editor sessions and delegation between persisted agent sessions. Scope: cli, ide Hooks / extensions Unknown Availability: Current primary documentation describes this product; installed behavior and maintenance cadence were not tested. [Source ↗](https://crow-ai.dev) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Crow Cli Readme ↗](https://github.com/crow-cli/crow-cli) official repository · checked 2026-09-21 [Crow Cli Acp Discovery ↗](https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json) registry · checked 2026-09-21 [Crow Cli License ↗](https://github.com/crow-cli/crow-cli/blob/main/LICENSE.md) official repository · checked 2026-09-21 | | | | |
| [Crush](#crush) Coding agent · Charmbracelet Terminal coding agent with model switching, language-server context and MCP integrations. | CLI | Multiple providers Local connection unknown | [FSL-1.1-MIT ↗](https://github.com/charmbracelet/crush/blob/main/LICENSE.md) Hyper plan, BYOK or local compute | Compare Crush Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Crush Charmbracelet mark · Profile reviewed 2026-09-21 Why consider Crush Use a terminal-focused interface with LSP context. Switch compatible models while preserving session context. What to weigh up Source is available under FSL with a competing-use restriction; it is not an unrestricted open-source license. Hooks are explicitly described as preliminary. Models & access Supports multiple providers and custom OpenAI- or Anthropic-compatible APIs. Local inference depends on endpoint compatibility and was not independently exercised. Hyper plan, BYOK or local compute Optional Charm Hyper offers a free tier and paid subscriptions. Alternatively, connect your own provider API keys or compatible local models. External inference and local compute costs depend on that choice; the client's source-available license is a separate consideration. [Check current plans and usage terms ↗](https://github.com/charmbracelet/crush) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/charmbracelet/crush) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/charmbracelet/crush) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Not open source ↗](https://github.com/charmbracelet/crush/blob/main/LICENSE.md) Source-available under FSL-1.1-MIT with a competing-use restriction; this is not unrestricted open source. Multiple model providers [Multiple providers ↗](https://github.com/charmbracelet/crush) Local model connection Unknown MCP [Documented ↗](https://github.com/charmbracelet/crush) Agent skills [Documented ↗](https://github.com/charmbracelet/crush) Subagents Unknown Hooks / extensions [Documented ↗](https://github.com/charmbracelet/crush) Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/charmbracelet/crush) Skills installation target: .crush/skills/ · Global: ~/.config/crush/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Crush README ↗](https://github.com/charmbracelet/crush) official repository · checked 2026-09-21 [Functional Source License ↗](https://github.com/charmbracelet/crush/blob/main/LICENSE.md) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Cursor](#cursor) Coding editor · Cursor Coding editor and agent platform with local, terminal and cloud workflows. | IDE CLI web | Multiple providers Local connection unknown | Licence not established Plans and metered model usage | Compare Cursor Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Cursor Product mark · Profile reviewed 2026-09-21 Why consider Cursor Keep code review, planning, editing and agent sessions together. Customize the agent with documented rules, skills, hooks and MCP. What to weigh up Model and usage allowances depend on the chosen plan. Cloud execution and editor execution have different data and environment boundaries. Models & access The managed catalogue spans multiple providers. Local-weight execution was not verified. A model selector is different from arbitrary local endpoints. Plans and metered model usage Use the current model/plan table; no fixed cost per task is implied. [Check current plans and usage terms ↗](https://cursor.com/docs/models-and-pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://cursor.com/downloads) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://cursor.com/docs) Editor integration [Documented ↗](https://cursor.com/docs) Desktop app Unknown Web interface [Documented ↗](https://cursor.com/docs) Open source Unknown Multiple model providers [Multiple providers ↗](https://cursor.com/docs/models-and-pricing) Hosted model selector in editor settings; arbitrary endpoints and all cloud/CLI parity are unverified. Scope: ide Local model connection Unknown MCP [Documented ↗](https://cursor.com/docs) Editor capability is verified; this profile does not infer CLI or web parity. Scope: ide Agent skills [Documented ↗](https://cursor.com/docs) Scope: ide Subagents [Documented ↗](https://cursor.com/docs) Scope: ide Hooks / extensions [Documented ↗](https://cursor.com/docs) Scope: ide Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://cursor.com/docs) Skills installation target: .agents/skills/ · Global: ~/.cursor/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Cursor documentation ↗](https://cursor.com/docs) official docs · checked 2026-09-21 [Models and pricing ↗](https://cursor.com/docs/models-and-pricing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Deep Agents Code](#deepagents) Coding agent · LangChain Terminal coding agent built on the Deep Agents SDK, with persistent memory, approval controls and configurable models. Also known as Deep Agents · dcode | CLI | Multiple providers Local connection documented | [MIT ↗](https://github.com/langchain-ai/deepagents/blob/main/LICENSE) Open-source client; model access separate | Compare Deep Agents Code Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Deep Agents Code LangChain mark · Profile reviewed 2026-09-21 Why consider Deep Agents Code Terminal coding with memory, subagents and model choice. Using configurable tool backends or remote sandboxes. What to weigh up The coding product and Python/JavaScript SDKs are distinct surfaces. Windows-specific installation was not established in this pass. Models & access Official SDK README explicitly supports tool-calling frontier APIs and self-hosted models via Ollama, vLLM or llama.cpp; Code docs say models/providers can be switched. Open-source client; model access separate Provider and optional hosted observability/sandbox costs are not included by the client license. [Check current plans and usage terms ↗](https://docs.langchain.com/deepagents-code) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.langchain.com/deepagents-code) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://docs.langchain.com/deepagents-code) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/langchain-ai/deepagents) Multiple model providers [Multiple providers ↗](https://docs.langchain.com/deepagents-code) Local model connection [Documented ↗](https://docs.langchain.com/deepagents-code) MCP [Documented ↗](https://docs.langchain.com/deepagents-code) Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Subagents [Documented ↗](https://docs.langchain.com/deepagents-code) Hooks / extensions [Documented ↗](https://docs.langchain.com/deepagents-code) Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/langchain-ai/deepagents) Skills installation target: .agents/skills/ · Global: ~/.deepagents/agent/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Deepagents Code ↗](https://docs.langchain.com/deepagents-code) official docs · checked 2026-09-21 [Deepagents Readme ↗](https://github.com/langchain-ai/deepagents) official repository · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [MIT license ↗](https://github.com/langchain-ai/deepagents/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Devin CLI](#devin) Coding agent · Cognition Local coding agent with editor integrations and handoff to Devin cloud environments. Also known as Devin for Terminal | CLI IDE | Multiple providers Local connection unknown | Licence not established Plan quota plus optional usage credits | Compare Devin CLI Add |
| Reasons, limits, setup & sources 7 sources · 2026-09-21 Devin CLI Cognition mark · Profile reviewed 2026-09-21 Why consider Devin CLI Start on local code and hand longer work to Devin Cloud. Use managed model selection with team configuration. What to weigh up CLI and cloud Devin are separate tools. CLI currently does not inherit Devin Knowledge, Playbooks or Secrets. Models & access Offers a managed catalog from several model vendors plus Adaptive routing and Fusion pairings. Enterprise teams can restrict the catalog; arbitrary BYOK/local endpoints were not verified. Plan quota plus optional usage credits Paid self-serve plans include CLI usage within allowances shared across Devin sessions and desktop. On-demand credits extend usage; team plans pool those credits across members. Enterprise contracts instead use Agent Compute Unit billing. Review the selected seat and plan because entitlements differ. [Check current plans and usage terms ↗](https://docs.devin.ai/admin/billing/self-serve) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.devin.ai/cli) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://docs.devin.ai/cli) Editor integration [Documented ↗](https://docs.devin.ai/cli) Desktop app Unknown Web interface Unknown Open source Unknown The reviewed sources did not establish this capability for a specific interface. Multiple model providers [Multiple providers ↗](https://docs.devin.ai/cli/models) The model-access evidence in this profile covers the terminal client. Other listed product surfaces must be checked separately. Offers a managed catalog from several model vendors plus Adaptive routing and Fusion pairings. Enterprise teams can restrict the catalog; arbitrary BYOK/local endpoints were not verified. Scope: cli Local model connection Unknown The reviewed sources did not establish this capability for a specific interface. MCP [Documented ↗](https://docs.devin.ai/cli) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Agent skills [Documented ↗](https://docs.devin.ai/cli) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents [Documented ↗](https://docs.devin.ai/cli/subagents) Devin CLI can delegate foreground or background tasks to isolated subagents with their own profiles and permission settings. Scope: cli Hooks / extensions [Documented ↗](https://docs.devin.ai/cli/extensibility/hooks/lifecycle-hooks) Lifecycle hooks are documented for Devin CLI sessions and tool events; IDE/ACP parity was not established. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://docs.devin.ai/cli) Skills installation target: .devin/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Devin CLI quickstart ↗](https://docs.devin.ai/cli) official docs · checked 2026-09-21 [Available models ↗](https://docs.devin.ai/cli/models) official docs · checked 2026-09-21 [Devin CLI subagents ↗](https://docs.devin.ai/cli/subagents) official docs · checked 2026-09-21 [Devin CLI lifecycle hooks ↗](https://docs.devin.ai/cli/extensibility/hooks/lifecycle-hooks) official docs · checked 2026-09-21 [Devin self-serve plans and shared usage quota ↗](https://docs.devin.ai/admin/billing/self-serve) official docs · checked 2026-09-21 [Devin self-serve and enterprise billing models ↗](https://docs.devin.ai/admin/billing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Devin Desktop](#windsurf) renamed Coding editor · Cognition Coding IDE formerly called Windsurf, now using Devin Local as its primary local agent. Also known as Windsurf | IDE | Multiple providers Local connection unknown | Licence not established Free tier; paid subscriptions and extra model usage | Compare Devin Desktop Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Devin Desktop Cognition mark · Profile reviewed 2026-09-21 Why consider Devin Desktop Move existing Windsurf projects into the documented successor IDE. Use built-in review, browser previews and local agent work. What to weigh up A Devin account is required. Legacy Cascade and Devin Local have different model availability; extension compatibility also differs from VS Code. Models & access Hosted model choices include different model families. Some models require Devin Local and are unavailable in legacy Cascade. Local agent execution is not proof of local model inference. Free tier; paid subscriptions and extra model usage Devin Free limits agent quota and model access. Paid plans expand both and allow extra usage purchases. Team plans combine an organization fee with developer-seat charges; enterprise controls are separately contracted. [Check current plans and usage terms ↗](https://devin.ai/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.devin.ai/desktop/getting-started) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration [Documented ↗](https://docs.devin.ai/desktop/getting-started) Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers [Multiple providers ↗](https://docs.devin.ai/desktop/models) Scope: ide Local model connection Unknown MCP [Documented ↗](https://docs.devin.ai/desktop/getting-started) Scope: ide Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Windsurf installation documentation now names Devin Desktop and documents transitional packages. [Source ↗](https://docs.devin.ai/desktop/getting-started) Skills installation target: .windsurf/skills/ · Global: ~/.codeium/windsurf/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Devin Desktop setup and rename ↗](https://docs.devin.ai/desktop/getting-started) official docs · checked 2026-09-21 [Devin Desktop models ↗](https://docs.devin.ai/desktop/models) official docs · checked 2026-09-21 [Devin Desktop plans, quotas and additional usage ↗](https://devin.ai/pricing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Dexto](#dexto) Coding agent · Truffle AI Configurable agent harness with a ready-to-use coding agent in terminal and web interfaces. | CLI web api | Multiple providers Local connection documented | [Elastic-2.0 ↗](https://github.com/truffle-ai/dexto/blob/main/LICENSE) Source available; model access separate | Compare Dexto Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Dexto Product mark · Profile reviewed 2026-09-21 Why consider Dexto Coding through either a CLI or web UI. Building configurable agents with local models and MCP tools. What to weigh up Source is available under Elastic License 2.0, which restricts offering the software as a managed service. Official README labels the project beta. Models & access Hosted providers, gateways, Ollama and local GGUF models are documented. Source available; model access separate Model-provider charges and Elastic License terms apply; review the terms for your deployment. [Check current plans and usage terms ↗](https://github.com/truffle-ai/dexto) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/truffle-ai/dexto) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/truffle-ai/dexto) Editor integration Unknown Desktop app Unknown Web interface [Documented ↗](https://github.com/truffle-ai/dexto) Open source [Not open source ↗](https://github.com/truffle-ai/dexto/blob/main/LICENSE) Multiple model providers [Multiple providers ↗](https://github.com/truffle-ai/dexto) Scope: cli, web Local model connection [Documented ↗](https://github.com/truffle-ai/dexto) Scope: cli, web MCP [Documented ↗](https://github.com/truffle-ai/dexto) Scope: cli, web Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Scope: cli, web Subagents [Documented ↗](https://github.com/truffle-ai/dexto) The documented coding agent can spawn specialist agents from both its CLI and Web UI; embedded API parity is not asserted. Scope: cli, web Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/truffle-ai/dexto) Skills installation target: .agents/skills/ · Global: ~/.agents/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Dexto Readme ↗](https://github.com/truffle-ai/dexto) official repository · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Elastic-2.0 license ↗](https://github.com/truffle-ai/dexto/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [DimAgent](#dimagent) renamed Coding agent · DimAgent Multi-model coding agent with terminal, desktop and ACP editor interfaces, custom providers and local state. Also known as DimCode · dimcode · dim | CLI desktop IDE | Multiple providers Local connection documented | Licence not established BYOK or local models; Coding Plan coming soon | Compare DimAgent Add |
| Reasons, limits, setup & sources 6 sources · 2026-09-21 DimAgent Product mark · Profile reviewed 2026-09-21 Why consider DimAgent Using local or hosted models across a terminal and desktop coding workspace. Forking tasks to background agents with configurable models. What to weigh up Public GitHub repository is an issue tracker, not product source. Web search requires a separate Serper key; planned features should not be assumed shipped. Models & access Official docs explicitly support custom endpoints, Ollama and LM Studio in addition to hosted providers. BYOK or local models; Coding Plan coming soon Hosted model costs depend on your connected provider; local inference uses your own hardware. The Coding Plan is listed as coming soon. [Check current plans and usage terms ↗](https://www.dimcode.dev/en/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://dimcode.dev/en/download/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://dimcode.dev/en/docs/) Editor integration [Documented ↗](https://dimcode.dev/en/docs/) Desktop app [Documented ↗](https://dimcode.dev/en/docs/) Web interface Unknown Open source Unknown Multiple model providers [Multiple providers ↗](https://dimcode.dev/en/docs/) Scope: cli, desktop Local model connection [Documented ↗](https://dimcode.dev/en/docs/) Scope: cli, desktop MCP [Documented ↗](https://dimcode.dev/en/docs/) Scope: cli, desktop Agent skills [Documented ↗](https://dimcode.dev/en/docs/) Scope: cli, desktop Subagents [Documented ↗](https://dimcode.dev/en/docs/) Documented CLI and desktop runtime features include background subagents and configurable child-model selection; external ACP editor parity is not asserted. Scope: cli, desktop Hooks / extensions [Documented ↗](https://dimcode.dev/en/docs/) Command hooks are documented in the CLI and the shared desktop runtime; external ACP editor parity is not asserted. Scope: cli, desktop Availability: Official site now labels the product DimAgent while package, command/config paths and ACP registry retain DimCode names. [Source ↗](https://dimcode.dev/en/docs/) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Dimagent Docs ↗](https://dimcode.dev/en/docs/) official docs · checked 2026-09-21 [Dimagent Cli ↗](https://dimcode.dev/en/docs/cli/) official docs · checked 2026-09-21 [Dimagent Features ↗](https://www.dimcode.dev/en/) official docs · checked 2026-09-21 [Dimagent Download ↗](https://dimcode.dev/en/download/) official docs · checked 2026-09-21 [Dimagent Issue Repo ↗](https://github.com/archships/dimcode) official repository · checked 2026-09-21 [Dimagent Acp Discovery ↗](https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json) registry · checked 2026-09-21 | | | | |
| [Dirac](#dirac) Coding agent · Dirac Delta Labs Coding agent for VS Code, terminal and ACP clients, with structured code tools and concurrent subagents. | CLI IDE | Multiple providers Local connection unknown | [Apache-2.0 ↗](https://github.com/dirac-run/dirac/blob/master/LICENSE) Open-source client; provider access separate | Compare Dirac Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Dirac Product mark · Profile reviewed 2026-09-21 Why consider Dirac Long coding tasks with separately configured subagents and user steering. Code inspection and edits using syntax-tree and stable-anchor tools. What to weigh up Structural editing depends on parser/index coverage and requires verification for dynamic references. Custom tool creation and automatic approvals depend on explicitly selected permissions/configuration. Models & access README describes provider choice and bring-your-own credentials with shared configuration across interfaces; local weights were not explicitly verified. Open-source client; provider access separate Costs depend on the model provider you connect. [Check current plans and usage terms ↗](https://github.com/dirac-run/dirac) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/dirac-run/dirac#install-dirac) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/dirac-run/dirac) Editor integration [Documented ↗](https://github.com/dirac-run/dirac) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/dirac-run/dirac/blob/master/LICENSE) Scope: cli, ide Multiple model providers [Multiple providers ↗](https://github.com/dirac-run/dirac) Scope: cli, ide Local model connection Unknown MCP Unknown Agent skills [Documented ↗](https://github.com/dirac-run/dirac) Scope: cli, ide Subagents [Documented ↗](https://github.com/dirac-run/dirac) The shared Dirac runtime documents concurrent subagents across terminal and editor entry points. Scope: cli, ide Hooks / extensions [Documented ↗](https://github.com/dirac-run/dirac) Repository hooks are documented for Dirac tasks in the shared CLI/editor runtime. Scope: cli, ide Availability: Current primary documentation describes this product; installed behavior and maintenance cadence were not tested. [Source ↗](https://dirac.run) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Dirac Readme ↗](https://github.com/dirac-run/dirac) official repository · checked 2026-09-21 [Dirac Acp Discovery ↗](https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json) registry · checked 2026-09-21 [Dirac License ↗](https://github.com/dirac-run/dirac/blob/master/LICENSE) official repository · checked 2026-09-21 | | | | |
| [Droid](#droid) Coding agent · Factory AI Factory’s coding agent across terminal and app workflows, with configurable specialist agents and model providers. | CLI desktop web | Multiple providers Local connection documented | Licence not established Subscription allowance; extra usage and BYOK | Compare Droid Add |
| Reasons, limits, setup & sources 7 sources · 2026-09-21 Droid Product mark · Profile reviewed 2026-09-21 Why consider Droid Use company-specific models or local endpoints from CLI and desktop. Package team skills, hooks and specialist droids. What to weigh up Model flexibility is surface-specific. Factory warns that arbitrary custom models are not guaranteed to work out of the box. Models & access CLI and desktop support BYOK, generic compatible endpoints and local Ollama/vLLM-style servers. Custom models do not appear on Factory-hosted web/mobile surfaces. Subscription allowance; extra usage and BYOK Individual subscriptions include the app, CLI and SDK with rolling usage limits. Optional prepaid Extra Usage extends access. BYOK has an included allowance, then plan-specific Factory charges can apply in addition to the model provider's charges; it is not an unlimited free platform route. [Check current plans and usage terms ↗](https://docs.factory.ai/pricing/individuals) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.factory.ai/droid-cli/quickstart) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://docs.factory.ai/model-independence/byok) Editor integration Unknown Desktop app [Documented ↗](https://docs.factory.ai/model-independence/byok) Web interface [Documented ↗](https://docs.factory.ai/model-independence/byok) Open source Unknown The reviewed sources did not establish this capability for a specific interface. Multiple model providers [Multiple providers ↗](https://docs.factory.ai/model-independence/byok) Factory explicitly documents custom models in CLI and desktop through local settings.json. Those custom models do not appear on Factory-hosted web/mobile surfaces; arbitrary endpoints are not guaranteed to work. Scope: cli, desktop Local model connection [Documented ↗](https://docs.factory.ai/model-independence/byok) Factory explicitly documents custom models in CLI and desktop through local settings.json. Those custom models do not appear on Factory-hosted web/mobile surfaces; arbitrary endpoints are not guaranteed to work. Scope: cli, desktop MCP [Documented ↗](https://docs.factory.ai/llms.txt) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Agent skills [Documented ↗](https://docs.factory.ai/llms.txt) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents [Documented ↗](https://docs.factory.ai/harness/subagents) CLI custom droids run in separate contexts. They are non-interactive and cannot themselves spawn further subagents; organization policies still apply. Scope: cli Hooks / extensions [Documented ↗](https://docs.factory.ai/harness/hooks) The documented local Droid hook configuration runs commands at lifecycle and tool events. Hosted web and desktop parity was not established by this review. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://docs.factory.ai/model-independence/byok) Skills installation target: .agents/skills/ · Global: ~/.factory/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Custom models (BYOK) ↗](https://docs.factory.ai/model-independence/byok) official docs · checked 2026-09-21 [Official documentation index ↗](https://docs.factory.ai/llms.txt) official docs · checked 2026-09-21 [Current Droid quickstart ↗](https://docs.factory.ai/droid-cli/quickstart) official docs · checked 2026-09-21 [Factory custom droids and subagents ↗](https://docs.factory.ai/harness/subagents) official docs · checked 2026-09-21 [Factory harness hooks ↗](https://docs.factory.ai/harness/hooks) official docs · checked 2026-09-21 [Factory individual plan usage, Extra Usage and BYOK ↗](https://docs.factory.ai/pricing/individuals) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Eve](#eve) Framework / platform · Vercel Filesystem-oriented framework for building durable agents with tools, skills, channels and schedules. | CLI api | Not established Local connection unknown | [Apache-2.0 ↗](https://github.com/vercel/eve/blob/main/LICENSE) Framework; service costs separate | Compare Eve Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Eve Vercel mark · Profile reviewed 2026-09-21 Why consider Eve Building custom agents from files and typed tools. Adding schedules and channel integrations to an agent application. What to weigh up Framework and project scaffold rather than an end-user coding assistant. Official README states beta terms and possible API/behavior changes. Models & access AI Gateway model configuration is documented; arbitrary endpoint and local model support were not established. Framework; service costs separate Configured model services and deployment infrastructure have separate costs. [Check current plans and usage terms ↗](https://github.com/vercel/eve) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/vercel/eve) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/vercel/eve) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/vercel/eve/blob/main/LICENSE) Scope: cli, api Multiple model providers Unknown Local model connection Unknown MCP Unknown Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Scope: cli, api Subagents [Documented ↗](https://github.com/vercel/eve) Subagents are a documented capability of the framework runtime used by generated terminal and application agents; this is not a separate ready-made coding product. Scope: cli, api Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/vercel/eve) Skills installation target: agent/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Eve Readme ↗](https://github.com/vercel/eve) official repository · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Apache-2.0 license ↗](https://github.com/vercel/eve/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [fast-agent](#fast-agent) Coding agent · evalstate CLI-first agent and workflow toolkit with an interactive coding mode, MCP integration and ACP editor access. | CLI IDE | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/evalstate/fast-agent/blob/main/LICENSE) Open-source client; provider access separate | Compare fast-agent Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 fast-agent Product mark · Profile reviewed 2026-09-21 Why consider fast-agent Interactive coding with shell tools, skills and MCP servers. Composing agent workflows and using the same setup through ACP. What to weigh up Shell support and built-in subagents are enabled by CLI options; capabilities depend on configuration. Windows filesystem/Docker MCP setups need documented configuration adjustments. Models & access Supports hosted providers and generic local endpoints; README explicitly demonstrates local llama.cpp configuration. Open-source client; provider access separate Hosted model services and local inference hardware have separate costs. [Check current plans and usage terms ↗](https://github.com/evalstate/fast-agent) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/evalstate/fast-agent) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/evalstate/fast-agent) Editor integration [Documented ↗](https://github.com/evalstate/fast-agent) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/evalstate/fast-agent/blob/main/LICENSE) Scope: cli, ide Multiple model providers [Multiple providers ↗](https://github.com/evalstate/fast-agent) Scope: cli, ide Local model connection [Documented ↗](https://github.com/evalstate/fast-agent) Scope: cli, ide MCP [Documented ↗](https://github.com/evalstate/fast-agent) Scope: cli, ide Agent skills [Documented ↗](https://github.com/evalstate/fast-agent) Scope: cli, ide Subagents [Documented ↗](https://github.com/evalstate/fast-agent) The README documents built-in subagents and states any fast-agent setup or program can be used through ACP. Scope: cli, ide Hooks / extensions [Documented ↗](https://github.com/evalstate/fast-agent) Agent and tool hooks are documented for fast-agent setups, which the README explicitly makes available through ACP as well as the CLI. Scope: cli, ide Availability: Current primary documentation describes this product; installed behavior and maintenance cadence were not tested. [Source ↗](https://fast-agent.ai) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Fast Agent Readme ↗](https://github.com/evalstate/fast-agent) official repository · checked 2026-09-21 [Fast Agent Acp Discovery ↗](https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json) registry · checked 2026-09-21 [Fast Agent License ↗](https://github.com/evalstate/fast-agent/blob/main/LICENSE) official repository · checked 2026-09-21 | | | | |
| [Firebender](#firebender) Coding agent · Firebender Coding agent integrated with Android Studio and JetBrains IDE tools. | IDE | Multiple providers Local connection unknown | Licence not established Subscription plus model usage | Compare Firebender Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Firebender Product mark · Profile reviewed 2026-09-21 Why consider Firebender Use Android/JetBrains context, refactoring tools and Compose previews. Connect a team model gateway through documented LiteLLM configuration. What to weigh up BYOK and on-demand usage require an active paid plan. Some tiers add a tool fee alongside model inference costs. Models & access Hosted model selection, BYOK, Amazon Bedrock and organization-managed LiteLLM are documented. Generic on-device model support is not established by the gateway integration alone. Subscription plus model usage Included usage, extra usage, BYOK availability and fees differ by tier; bonus capacity is not guaranteed. [Check current plans and usage terms ↗](https://docs.firebender.com/get-started/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.firebender.com/get-started/quickstart) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration [Documented ↗](https://docs.firebender.com/get-started/welcome) Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers [Multiple providers ↗](https://docs.firebender.com/configurations/firebender-on-litellm) Scope: ide Local model connection Unknown MCP [Documented ↗](https://docs.firebender.com/get-started/welcome) Scope: ide Agent skills [Documented ↗](https://docs.firebender.com/get-started/welcome) Scope: ide Subagents [Documented ↗](https://docs.firebender.com/get-started/welcome) Scope: ide Hooks / extensions [Documented ↗](https://docs.firebender.com/get-started/welcome) Scope: ide Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://docs.firebender.com/get-started/welcome) Skills installation target: .agents/skills/ · Global: ~/.firebender/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Firebender documentation ↗](https://docs.firebender.com/get-started/welcome) official docs · checked 2026-09-21 [Firebender on LiteLLM ↗](https://docs.firebender.com/configurations/firebender-on-litellm) official docs · checked 2026-09-21 [Firebender pricing ↗](https://docs.firebender.com/get-started/pricing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [ForgeCode](#forgecode) Coding agent · Tailcall Open-source coding harness with shell integration, configurable agents and cloud or local model providers. Also known as Forge | CLI IDE | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/tailcallhq/forgecode/blob/main/LICENSE) Provider-funded; optional Forge services | Compare ForgeCode Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 ForgeCode Product mark · Profile reviewed 2026-09-21 Why consider ForgeCode Send coding tasks directly from a configured Zsh prompt. Choose cloud or local providers and tailor agent definitions. What to weigh up Zsh integration and terminal font setup add onboarding steps. Windows uses WSL or Git Bash in the documented install path. Models & access Official docs support cloud models, open-weight models and local models with provider configuration. Provider-funded; optional Forge services Connect a provider API key, supported provider account or local endpoint. Model usage and local compute follow that choice. Optional ForgeCode Services add hosted context and guardrail services; signing into those services is distinct from configuring the model provider. [Check current plans and usage terms ↗](https://forgecode.dev/docs/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://forgecode.dev/docs/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://forgecode.dev/docs/) Editor integration [Documented ↗](https://forgecode.dev/docs/) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/tailcallhq/forgecode/blob/main/LICENSE) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Multiple model providers [Multiple providers ↗](https://forgecode.dev/docs/) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Local model connection [Documented ↗](https://forgecode.dev/docs/) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli MCP [Documented ↗](https://forgecode.dev/docs/) Verified for the terminal agent. Listing an editor/ACP integration does not by itself establish every host editor’s MCP setup or parity. Scope: cli Agent skills [Documented ↗](https://forgecode.dev/docs/) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents Unknown Hooks / extensions Unknown Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://forgecode.dev/docs/) Skills installation target: .forge/skills/ · Global: ~/.forge/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Installation and setup ↗](https://forgecode.dev/docs/) official docs · checked 2026-09-21 [Repository ↗](https://github.com/tailcallhq/forgecode) official repository · checked 2026-09-21 [Apache license ↗](https://github.com/tailcallhq/forgecode/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [fx](#fx) Coding agent · Vercel Labs Native terminal coding agent with a shell-like interface and an embeddable agent core. | CLI | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/vercel-labs/fx/blob/main/LICENSE) Open-source client; model access separate | Compare fx Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 fx Product mark · Profile reviewed 2026-09-21 Why consider fx Shell-oriented terminal coding and one-shot script tasks. Embedding the same agent core through libfx or ACP. What to weigh up Official README labels fx experimental. Official install documentation lists macOS/Linux only; Windows support not established. Models & access Supports Vercel AI Gateway, supported subscriptions, and custom OpenAI-compatible endpoints including Ollama. Open-source client; model access separate Model gateway, provider or subscription access is separate from client license. [Check current plans and usage terms ↗](https://github.com/vercel-labs/fx) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://fx.sh/docs/getting-started/installation) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/vercel-labs/fx) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/vercel-labs/fx/blob/main/LICENSE) Multiple model providers [Multiple providers ↗](https://github.com/vercel-labs/fx) Local model connection [Documented ↗](https://github.com/vercel-labs/fx) MCP [Documented ↗](https://github.com/vercel-labs/fx) Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Subagents [Documented ↗](https://github.com/vercel-labs/fx) Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/vercel-labs/fx) Skills installation target: .fx/skills/ · Global: ~/.fx/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Fx Readme ↗](https://github.com/vercel-labs/fx) official repository · checked 2026-09-21 [Fx Install ↗](https://fx.sh/docs/getting-started/installation) official docs · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Apache-2.0 license ↗](https://github.com/vercel-labs/fx/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Gemini CLI](#gemini-cli) Coding agent · Google Google’s open-source terminal agent for coding, local files, shell automation and Google-backed model access. | CLI IDE | Provider ecosystem Local connection unknown | [Apache-2.0 ↗](https://github.com/google-gemini/gemini-cli/blob/main/LICENSE) Free allowance, subscription or metered API | Compare Gemini CLI Add |
| Reasons, limits, setup & sources 6 sources · 2026-09-21 Gemini CLI Product mark · Profile reviewed 2026-09-21 Why consider Gemini CLI Use Gemini from a terminal with shell, search and MCP tools. Extend an open-source CLI with skills, hooks and editor integration. What to weigh up Some advanced capabilities are marked experimental in the docs. Google authentication tier and API route determine available usage. Models & access Documented primary model and authentication paths use Gemini. This pass did not verify arbitrary local inference support. Free allowance, subscription or metered API Billing depends on authentication. Personal Google sign-in has an included allowance; eligible Google AI and Gemini Code Assist subscriptions offer higher limits. Gemini API keys and Vertex AI have their own free or usage-based tiers. CLI requests and tool work consume the selected route's quota. [Check current plans and usage terms ↗](https://geminicli.com/docs/resources/quota-and-pricing/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://geminicli.com/docs/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/google-gemini/gemini-cli) Editor integration [Documented ↗](https://github.com/google-gemini/gemini-cli) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/google-gemini/gemini-cli) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Multiple model providers [One provider ecosystem ↗](https://github.com/google-gemini/gemini-cli) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Local model connection Unknown The reviewed sources did not establish this capability for a specific interface. MCP [Documented ↗](https://geminicli.com/docs/) Verified for the terminal agent. Listing an editor/ACP integration does not by itself establish every host editor’s MCP setup or parity. Scope: cli Agent skills [Documented ↗](https://geminicli.com/docs/) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents [Documented ↗](https://geminicli.com/docs/core/subagents/) The current CLI documentation enables subagents by default and exposes management through /agents. Editor/ACP integration does not independently establish feature parity. Scope: cli Hooks / extensions [Documented ↗](https://geminicli.com/docs/hooks/) CLI lifecycle hooks run synchronously during the agent loop. Their configuration and event format are specific to Gemini CLI. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/google-gemini/gemini-cli) Skills installation target: .agents/skills/ · Global: ~/.gemini/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Gemini CLI README ↗](https://github.com/google-gemini/gemini-cli) official repository · checked 2026-09-21 [Documentation feature index ↗](https://geminicli.com/docs/) official docs · checked 2026-09-21 [Gemini CLI subagents ↗](https://geminicli.com/docs/core/subagents/) official docs · checked 2026-09-21 [Gemini CLI hooks ↗](https://geminicli.com/docs/hooks/) official docs · checked 2026-09-21 [Gemini CLI quotas and pricing by authentication method ↗](https://geminicli.com/docs/resources/quota-and-pricing/) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [GitHub Copilot CLI](#github-copilot) Coding agent · GitHub GitHub-integrated terminal agent for local code and repository, issue and pull-request workflows. | CLI IDE | Multiple providers Local connection unknown | [proprietary ↗](https://github.com/github/copilot-cli/blob/main/LICENSE.md) Copilot plan usage | Compare GitHub Copilot CLI Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 GitHub Copilot CLI Product mark · Profile reviewed 2026-09-21 Why consider GitHub Copilot CLI Work with GitHub issues and pull requests from the terminal. Use Copilot plans and organization policy across local and cloud agent tasks. What to weigh up Organization users need the Copilot CLI policy enabled. Local/cloud sandbox features are documented as preview. Models & access Managed model selection spans several vendors. This is not evidence of arbitrary local endpoints; organization policies can constrain access. Copilot plan usage Available with Copilot plans; organization policy and plan limits apply. [Check current plans and usage terms ↗](https://docs.github.com/en/copilot/concepts/agents/copilot-cli/about-copilot-cli) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/github/copilot-cli) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://docs.github.com/en/copilot/concepts/agents/copilot-cli/about-copilot-cli) Editor integration [Documented ↗](https://docs.github.com/en/copilot/concepts/agents/copilot-cli/about-copilot-cli) Desktop app Unknown Web interface Unknown Open source [Not open source ↗](https://github.com/github/copilot-cli/blob/main/LICENSE.md) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Multiple model providers [Multiple providers ↗](https://docs.github.com/en/copilot/concepts/agents/copilot-cli/about-copilot-cli) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Local model connection Unknown The reviewed sources did not establish this capability for a specific interface. MCP [Documented ↗](https://docs.github.com/en/copilot/concepts/agents/copilot-cli/about-copilot-cli) Verified for the terminal agent. Listing an editor/ACP integration does not by itself establish every host editor’s MCP setup or parity. Scope: cli Agent skills [Documented ↗](https://docs.github.com/en/copilot/concepts/agents/copilot-cli/about-copilot-cli) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents [Documented ↗](https://docs.github.com/en/copilot/concepts/agents/copilot-cli/about-copilot-cli) The CLI documentation describes built-in specialized agents and custom agents for delegated work. This profile's ACP integration does not imply every IDE feature is equivalent. Scope: cli Hooks / extensions [Documented ↗](https://docs.github.com/en/copilot/concepts/agents/copilot-cli/about-copilot-cli) Copilot CLI supports custom shell commands at agent lifecycle points. The CLI documentation is the scope evidence; broader Copilot editor behavior is separate. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://docs.github.com/en/copilot/concepts/agents/copilot-cli/about-copilot-cli) Skills installation target: .agents/skills/ · Global: ~/.copilot/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [About Copilot CLI ↗](https://docs.github.com/en/copilot/concepts/agents/copilot-cli/about-copilot-cli) official repository · checked 2026-09-21 [Copilot CLI README ↗](https://github.com/github/copilot-cli) official repository · checked 2026-09-21 [Official license ↗](https://github.com/github/copilot-cli/blob/main/LICENSE.md) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [goose](#goose) AI assistant · Agentic AI Foundation Open-source general-purpose agent with a desktop app, terminal client, model-provider choices and MCP extensions. | CLI desktop | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/aaif-goose/goose/blob/main/LICENSE) Provider APIs, linked plans or local compute | Compare goose Add |
| Reasons, limits, setup & sources 7 sources · 2026-09-21 goose Product mark · Profile reviewed 2026-09-21 Why consider goose Use the same general-purpose agent from desktop or terminal. Connect local models and MCP tools for research, files and development. What to weigh up Model setup and extension configuration remain your responsibility. A desktop install and CLI install are separate choices. Models & access Supports hosted API providers and Ollama. ACP integrations can use supported existing coding subscriptions; inference terms remain with the provider. Provider APIs, linked plans or local compute Configure a hosted provider API key, a supported subscription-backed integration or a local model runner. Goose does not supply a universal model allowance: billing and limits follow the selected provider or linked account. Local models use your own machine or server resources. [Check current plans and usage terms ↗](https://goose-docs.ai/docs/getting-started/providers/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://goose-docs.ai/docs/getting-started/installation) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/aaif-goose/goose) Editor integration Unknown Desktop app [Documented ↗](https://github.com/aaif-goose/goose) Web interface Unknown Open source [Open source ↗](https://github.com/aaif-goose/goose) Documented for the Goose CLI and desktop application. Provider/model compatibility and extension configuration still apply to each session. Scope: cli, desktop Multiple model providers [Multiple providers ↗](https://github.com/aaif-goose/goose) Documented for the Goose CLI and desktop application. Provider/model compatibility and extension configuration still apply to each session. Scope: cli, desktop Local model connection [Documented ↗](https://github.com/aaif-goose/goose) Documented for the Goose CLI and desktop application. Provider/model compatibility and extension configuration still apply to each session. Scope: cli, desktop MCP [Documented ↗](https://github.com/aaif-goose/goose) Documented for the Goose CLI and desktop application. Provider/model compatibility and extension configuration still apply to each session. Scope: cli, desktop Agent skills [Documented ↗](https://goose-docs.ai/docs/guides/context-engineering/using-skills) Documented for the Goose CLI and desktop application. Provider/model compatibility and extension configuration still apply to each session. Scope: cli, desktop Subagents [Documented ↗](https://goose-docs.ai/docs/guides/context-engineering/subagents) Subagent activity is documented for desktop and CLI. Subagents require autonomous permission mode and are disabled in manual approval, smart approval and chat-only modes. Scope: cli, desktop Hooks / extensions Unknown Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/aaif-goose/goose) Skills installation target: .goose/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [goose README ↗](https://github.com/aaif-goose/goose) official repository · checked 2026-09-21 [Installation ↗](https://goose-docs.ai/docs/getting-started/installation) official docs · checked 2026-09-21 [Skills ↗](https://goose-docs.ai/docs/guides/context-engineering/using-skills) official docs · checked 2026-09-21 [Subagents ↗](https://goose-docs.ai/docs/guides/context-engineering/subagents) official docs · checked 2026-09-21 [CLI and desktop quickstart ↗](https://goose-docs.ai/docs/quickstart/) official docs · checked 2026-09-21 [Configure providers in desktop and CLI ↗](https://goose-docs.ai/docs/getting-started/providers/) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [GPT4All](#gpt4all) Model client · Nomic Local desktop model runner with document chat and a Python interface. | desktop | Multiple providers Local connection documented | [MIT ↗](https://github.com/nomic-ai/gpt4all/blob/main/LICENSE.txt) Open-source client; local inference | Compare GPT4All Add |
| Reasons, limits, setup & sources 2 sources · 2026-09-21 GPT4All Nomic mark · Profile reviewed 2026-09-21 Why consider GPT4All Start local document chat without a provider account. Use a desktop client or integrate local inference through Python. What to weigh up Available model capabilities and hardware still constrain results. A local chat client is not a drop-in autonomous coding agent. Models & access Runs downloadable local models on desktops; the project also exposes a Python client. No API key is required for local model execution. Open-source client; local inference Model licenses and hardware costs remain separate. [Check current plans and usage terms ↗](https://github.com/nomic-ai/gpt4all) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.gpt4all.io/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://github.com/nomic-ai/gpt4all) Web interface Unknown Open source [Open source ↗](https://github.com/nomic-ai/gpt4all) Scope: desktop Multiple model providers [Multiple providers ↗](https://docs.gpt4all.io/gpt4all_desktop/models.html) The desktop Models guide explicitly supports adding API keys for remote model providers. This is separate from downloaded model weights: API prompts leave the computer for the selected provider. Scope: desktop Local model connection [Documented ↗](https://github.com/nomic-ai/gpt4all) Scope: desktop MCP Unknown Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://github.com/nomic-ai/gpt4all) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [GPT4All official repository ↗](https://github.com/nomic-ai/gpt4all) official repository · checked 2026-09-21 [GPT4All desktop local models and remote model APIs ↗](https://docs.gpt4all.io/gpt4all_desktop/models.html) official docs · checked 2026-09-21 | | | | |
| [Grok Build](#grok) Coding agent · xAI Open-source Rust CLI/TUI coding agent from xAI, also usable headlessly and through ACP. Also known as Grok CLI | CLI IDE | Multiple providers Local connection unknown | [Apache-2.0 ↗](https://github.com/xai-org/grok-build/blob/main/LICENSE) Metered API or configured provider | Compare Grok Build Add |
| Reasons, limits, setup & sources 7 sources · 2026-09-21 Grok Build Product mark · Profile reviewed 2026-09-21 Why consider Grok Build Use the terminal TUI or headless CLI with a configured model ID and API base URL. Delegate repository exploration and planning to separate child contexts with built-in types that cannot run shell commands or edit files. Extend the terminal workflow with skills, plugin marketplaces, MCP servers and lifecycle hooks. Inspect and customize the published Apache-2.0 CLI/runtime source. What to weigh up Sandboxing is off by default. Documented isolation uses Landlock on Linux and Seatbelt on macOS; enable a profile before relying on filesystem limits. Child-process network restrictions apply on Linux only; they do not block model API calls or built-in web tools. Built-in sandbox profiles do not permanently protect credential paths such as ~/.ssh; the docs advise a custom deny list. Session storage remains writable. Skill allowed-tools metadata does not enforce tool access, and general-purpose subagents retain full capabilities. The public repository is periodically synchronized from the vendor’s internal repository. The open-source client licence does not include hosted inference; account and provider billing still apply. Models & access The getting-started guide demonstrates grok-4.7 on the xAI API and documents custom model IDs and base URLs. It does not establish a universal Grok Build default. Local model serving was not separately verified. Metered API or configured provider The xAI API-key route uses prepaid API credits or approved invoiced billing. Custom model endpoints use their own provider accounts and terms. Browser sign-in is also documented, but this review did not establish a bundled subscription allowance for that route. [Check current plans and usage terms ↗](https://docs.x.ai/console/billing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.x.ai/build/overview) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/xai-org/grok-build) Editor integration [Documented ↗](https://github.com/xai-org/grok-build) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/xai-org/grok-build) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Multiple model providers [Multiple providers ↗](https://docs.x.ai/build/overview) The CLI accepts a configured model ID and API base URL. This establishes configurable endpoints, not compatibility with every provider or local model server; editor/ACP parity was not separately verified. Scope: cli Local model connection Unknown The reviewed sources did not establish this capability for a specific interface. MCP [Documented ↗](https://github.com/xai-org/grok-build) Verified for the terminal agent. Listing an editor/ACP integration does not by itself establish every host editor’s MCP setup or parity. Scope: cli Agent skills [Documented ↗](https://docs.x.ai/build/features/skills-plugins-marketplaces) The terminal client discovers project, user and plugin skills. SKILL.md allowed-tools metadata does not grant or restrict tools. Permissions decide whether a tool call runs; a sandbox profile limits what an approved call can do. Editor/ACP parity was not separately verified. Scope: cli Subagents [Documented ↗](https://docs.x.ai/build/features/subagents) Independent CLI child sessions have separate context and return summaries to the parent; subagents are enabled by default when the setting is unset. Built-in explore reads, lists and searches, while explore and plan cannot run shell commands or edit files. General-purpose children retain full capabilities. Editor/ACP parity was not separately verified. Scope: cli Hooks / extensions [Documented ↗](https://github.com/xai-org/grok-build) CLI/TUI hooks run on tool and session lifecycle events; project hooks require explicit trust through /hooks-trust. Editor/ACP parity was not separately verified. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/xai-org/grok-build) Skills installation target: .grok/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Grok Build source README ↗](https://github.com/xai-org/grok-build) official repository · checked 2026-09-21 [Grok Build overview ↗](https://docs.x.ai/build/overview) official docs · checked 2026-09-21 [xAI API credits and invoiced billing ↗](https://docs.x.ai/console/billing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Grok Build independent child sessions and built-in agent types ↗](https://docs.x.ai/build/features/subagents) official docs · checked 2026-09-21 [Grok Build sandbox profiles and platform limits ↗](https://docs.x.ai/build/features/sandbox) official docs · checked 2026-09-21 [Grok Build skills, plugins, marketplaces and permission metadata ↗](https://docs.x.ai/build/features/skills-plugins-marketplaces) official docs · checked 2026-09-21 | | | | |
| [Harn](#harn) Framework / platform · Burin Labs Programming language and runtime for building agents with tools, permissions, recorded execution and child agents. | CLI | Multiple providers Local connection documented | [MIT OR Apache-2.0 ↗](https://github.com/burin-labs/harn#license) Open-source runtime; model costs separate | Compare Harn Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Harn Product mark · Profile reviewed 2026-09-21 Why consider Harn Building custom agents whose tool permissions and execution history are part of application code. Testing workflows with recorded model responses and deterministic replay. What to weigh up A programming language/runtime, not a ready-made coding assistant. Pre-release language and CLI can change between releases. Models & access Primary README and getting-started guide explicitly support hosted and local models, including Ollama/llama.cpp/MLX/vLLM routes. Open-source runtime; model costs separate Offline recorded-response demos require no provider; live hosted calls require credentials and may incur charges. [Check current plans and usage terms ↗](https://github.com/burin-labs/harn) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/burin-labs/harn/blob/main/docs/src/getting-started.md) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/burin-labs/harn) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/burin-labs/harn#license) Multiple model providers [Multiple providers ↗](https://github.com/burin-labs/harn) Local model connection [Documented ↗](https://github.com/burin-labs/harn) MCP [Documented ↗](https://harnlang.com) Agent skills Unknown Subagents [Documented ↗](https://github.com/burin-labs/harn) Hooks / extensions Unknown Availability: Current primary documentation describes this product; installed behavior and maintenance cadence were not tested. [Source ↗](https://harnlang.com) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Harn Readme ↗](https://github.com/burin-labs/harn) official repository · checked 2026-09-21 [Harn Install ↗](https://github.com/burin-labs/harn/blob/main/docs/src/getting-started.md) official repository · checked 2026-09-21 [Harn Protocols ↗](https://harnlang.com) official docs · checked 2026-09-21 [Harn Acp Discovery ↗](https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json) registry · checked 2026-09-21 [Harn License ↗](https://github.com/burin-labs/harn#license) official repository · checked 2026-09-21 | | | | |
| [Hermes Agent](#hermes-agent) AI assistant · Nous Research General-purpose agent combining terminal work, persistent memory, reusable skills and messaging integrations. | CLI desktop | Multiple providers Local connection documented | [MIT ↗](https://github.com/NousResearch/hermes-agent/blob/main/LICENSE) Optional Nous plan, BYOK or local compute | Compare Hermes Agent Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Hermes Agent Product mark · Profile reviewed 2026-09-21 Why consider Hermes Agent Run recurring research and automation with persistent memories and skills. Continue work through supported messaging channels and remote execution backends. What to weigh up Persistent memory and remote tools need intentional configuration. External models, hosting and optional tool services have their own costs. Models & access Supports Nous Portal, external APIs, local Ollama/vLLM and compatible endpoints. The local server must meet the tool-use context requirements documented by Hermes. Optional Nous plan, BYOK or local compute Run the agent with external API keys or a locally served model, paying the chosen provider and any hosting costs. An optional Nous Portal subscription bundles model access and supported tool services. Separately connected search, browser or other tool accounts retain their own terms. [Check current plans and usage terms ↗](https://hermes-agent.nousresearch.com/docs/integrations/providers) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/NousResearch/hermes-agent) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/NousResearch/hermes-agent) Editor integration Unknown Desktop app [Documented ↗](https://github.com/NousResearch/hermes-agent) Web interface Unknown Open source [Open source ↗](https://github.com/NousResearch/hermes-agent) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Multiple model providers [Multiple providers ↗](https://github.com/NousResearch/hermes-agent) The model-access evidence in this profile covers the terminal client. Other listed product surfaces must be checked separately. Supports Nous Portal, external APIs, local Ollama/vLLM and compatible endpoints. The local server must meet the tool-use context requirements documented by Hermes. Scope: cli Local model connection [Documented ↗](https://github.com/NousResearch/hermes-agent) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli MCP [Documented ↗](https://github.com/NousResearch/hermes-agent) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Agent skills [Documented ↗](https://github.com/NousResearch/hermes-agent) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents [Documented ↗](https://github.com/NousResearch/hermes-agent) The agent repository documents isolated subagents for parallel workstreams. This audit establishes the CLI agent capability, not feature parity in a desktop wrapper. Scope: cli Hooks / extensions Unknown Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/NousResearch/hermes-agent) Skills installation target: .hermes/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Hermes Agent README ↗](https://github.com/NousResearch/hermes-agent) official repository · checked 2026-09-21 [Model providers ↗](https://hermes-agent.nousresearch.com/docs/integrations/providers) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [IBM Bob](#bob) Coding editor · IBM IBM coding and modernization assistant with Ask, Plan and Agent modes and parallel subagents. | IDE CLI | Not established Local connection unknown | Licence not established Subscription with Bobcoin allowance; limited free trial | Compare IBM Bob Add |
| Reasons, limits, setup & sources 7 sources · 2026-09-21 IBM Bob Product mark · Profile reviewed 2026-09-21 Why consider IBM Bob Review an implementation plan before allowing edits. Use explicit subagent approvals and configurable external tools. What to weigh up Old IDE releases are scheduled to stop working; use the current official installer. MCP servers are not preinstalled and require configuration. Models & access Bob uses managed model orchestration. The fetched FAQ rendered questions without answers, so model choice and local inference remain unverified. Subscription with Bobcoin allowance; limited free trial Individual tiers include Bobcoins and Bob Shell access. Enterprise adds centralized user and coin management. Specialist modernization packages are separate paid add-ons, and their trials require a subscription. [Check current plans and usage terms ↗](https://bob.ibm.com/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://bob.ibm.com/docs/ide) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://bob.ibm.com/docs/ide) Editor integration [Documented ↗](https://bob.ibm.com/docs/ide) Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP [Documented ↗](https://bob.ibm.com/docs/ide/configuration/mcp/mcp-in-bob) Verified for Bob IDE. Bob Shell parity was not established from the reviewed pages. Scope: ide Agent skills [Documented ↗](https://bob.ibm.com/docs/ide) Scope: ide Subagents [Documented ↗](https://bob.ibm.com/docs/ide) Scope: ide Hooks / extensions [Documented ↗](https://bob.ibm.com/docs/shell/features/slash-commands) IDE hook support and Shell hook-management command are separately documented. Scope: ide, cli Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://bob.ibm.com/docs/ide) Skills installation target: .bob/skills/ · Global: ~/.bob/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [IBM Bob overview ↗](https://bob.ibm.com/docs/ide) official docs · checked 2026-09-21 [Using MCP in Bob ↗](https://bob.ibm.com/docs/ide/configuration/mcp/mcp-in-bob) official docs · checked 2026-09-21 [IBM Bob FAQ ↗](https://bob.ibm.com/docs/ide/faq) official docs · checked 2026-09-21 [Bob Shell overview ↗](https://bob.ibm.com/docs/shell) official docs · checked 2026-09-21 [Bob Shell slash commands ↗](https://bob.ibm.com/docs/shell/features/slash-commands) official docs · checked 2026-09-21 [IBM Bob subscriptions, trial and premium packages ↗](https://bob.ibm.com/pricing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [iFlow CLI (retired)](#iflow-cli) retired Coding agent · iFlow Historical terminal coding assistant whose maintainers announced service shutdown. | CLI IDE | Not established Local connection unknown | Licence not established Retired; former free access unavailable | Compare iFlow CLI (retired) Add |
| Reasons, limits, setup & sources 2 sources · 2026-09-21 iFlow CLI (retired) iFlow mark · Profile reviewed 2026-09-21 Why consider iFlow CLI (retired) Identify a legacy installation and locate migration guidance. What to weigh up Official README announces shutdown on April 17, 2026 (UTC+8). Do not recommend new installations or advertise former free-model access. Models & access Historical provider access is not evidence of current availability after the announced shutdown. Retired; former free access unavailable The official repository announces the service shutdown. Historical free-access claims and old usage allowances should not be treated as an available offer. Review the shutdown and migration information before relying on an existing installation. [Check current plans and usage terms ↗](https://github.com/iflow-ai/iflow-cli) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/iflow-ai/iflow-cli) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/iflow-ai/iflow-cli) Editor integration [Documented ↗](https://github.com/iflow-ai/iflow-cli) Desktop app Unknown Web interface Unknown Open source Unknown The reviewed sources did not establish this capability for a specific interface. Multiple model providers Unknown The reviewed sources did not establish this capability for a specific interface. Local model connection Unknown The reviewed sources did not establish this capability for a specific interface. MCP [Documented ↗](https://github.com/iflow-ai/iflow-cli) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Agent skills Unknown The reviewed sources did not establish this capability for a specific interface. Subagents [Documented ↗](https://github.com/iflow-ai/iflow-cli) The retired CLI's README documents historical subagent support. Its shutdown notice takes precedence over these historical capabilities for current recommendations. Scope: cli Hooks / extensions [Documented ↗](https://github.com/iflow-ai/iflow-cli) The retired CLI's README lists historical hook support. This is not evidence that the shut-down service remains available. Scope: cli Availability: Official README opens with the April 17, 2026 shutdown notice; retain only as retired/migration information, not a current tool recommendation. [Source ↗](https://github.com/iflow-ai/iflow-cli) Skills installation target: .iflow/skills/ · Global: ~/.iflow/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Official shutdown notice and historical README ↗](https://github.com/iflow-ai/iflow-cli) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [inference.sh / Belt](#inference-sh) Framework / platform · inference.sh Cloud agent and AI-tool platform with the Belt CLI for skills, tools, connectors and deployment. Also known as Belt · belt · infsh · inferencesh | CLI web api | Not established Local connection unknown | Licence not established Cloud usage/account billing | Compare inference.sh / Belt Add |
| Reasons, limits, setup & sources 6 sources · 2026-09-21 inference.sh / Belt Product mark · Profile reviewed 2026-09-21 Why consider inference.sh / Belt Calling hosted AI tools and deploying durable agents. Distributing skills and connecting existing coding-agent harnesses. What to weigh up Cloud platform account and usage limits apply. Belt orchestration and tool access should not be presented as a standalone local coding assistant. Models & access Cloud model catalogue and BYOK are described; this does not establish arbitrary endpoints or local weights. Cloud usage/account billing Your platform plan, usage and connected model providers determine costs; check current account terms. [Check current plans and usage terms ↗](https://inference.sh/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://inference.sh/docs/extend/cli-setup) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://inference.sh) Editor integration Unknown Desktop app Unknown Web interface [Documented ↗](https://inference.sh) Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP [Documented ↗](https://inference.sh/docs/extend/cli-setup) Scope: api Agent skills [Documented ↗](https://inference.sh/docs/extend/cli-setup) Scope: cli, api Subagents [Documented ↗](https://inference.sh/docs/extend/cli-setup) Official sub-agent documentation establishes delegation configured through the web agent settings Tools tab. CLI/API interface parity was not separately established. Scope: web Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://inference.sh) Skills installation target: .inferencesh/skills/ · Global: ~/.inferencesh/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Inference Sh Primary ↗](https://inference.sh) official docs · checked 2026-09-21 [Inference Cli ↗](https://inference.sh/docs/extend/cli-setup) official docs · checked 2026-09-21 [Inference Platform ↗](https://inference.sh/docs/extend/cli-setup) official docs · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [inference.sh sub-agents ↗](https://inference.sh/docs/agents/sub-agents) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Jan Agent](#jan-agent) Coding agent · Jan Preview terminal coding agent sharing model providers and extensions with Jan Desktop. | CLI TUI | Multiple providers Local connection documented | Licence not established Bring your own inference; preview client fee unverified | Compare Jan Agent Add |
| Reasons, limits, setup & sources 2 sources · 2026-09-21 Jan Agent Product mark · Profile reviewed 2026-09-21 Why consider Jan Agent Run folder-based file and shell work using an existing Jan model setup. Reuse skills, MCP and provider settings with Jan Desktop. What to weigh up Official docs label this a preview and the dev installer uses nightly builds. Local model tool reliability is still model-dependent. Models & access Uses cloud providers or OpenAI-compatible endpoints, including a local model served by Jan Desktop. Bring your own inference; preview client fee unverified Uses configured cloud-provider credentials or your own compatible model server, including Jan Desktop. No separate Agent client fee was stated in the reviewed official overview/provider docs; provider-account charges and local hardware requirements remain separate. [Check current plans and usage terms ↗](https://www.jan.ai/docs/agent/providers) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://www.jan.ai/docs/agent) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://www.jan.ai/docs/agent) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers [Multiple providers ↗](https://www.jan.ai/docs/agent/providers) The terminal agent accepts configured OpenAI-compatible and Anthropic-compatible provider endpoints and credentials. It can also use a separately served local model; the agent itself ships no inference engine. Scope: cli, tui Local model connection [Documented ↗](https://www.jan.ai/docs/agent) Scope: cli, tui MCP [Documented ↗](https://www.jan.ai/docs/agent) Scope: cli, tui Agent skills [Documented ↗](https://www.jan.ai/docs/agent) Scope: cli, tui Subagents [Documented ↗](https://www.jan.ai/docs/agent) Scope: cli, tui Hooks / extensions [Documented ↗](https://www.jan.ai/docs/agent) Scope: cli, tui Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://www.jan.ai/docs/agent) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Jan Agent overview and installation ↗](https://www.jan.ai/docs/agent) official docs · checked 2026-09-21 [Jan Agent provider credentials and own-hardware inference ↗](https://www.jan.ai/docs/agent/providers) official docs · checked 2026-09-21 | | | | |
| [Jan Desktop](#jan) Model client · Jan Open-source desktop client for local and cloud models with MCP and a local API server. | desktop | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/janhq/jan/blob/main/LICENSE) Open-source client; inference separate | Compare Jan Desktop Add |
| Reasons, limits, setup & sources 2 sources · 2026-09-21 Jan Desktop Product mark · Profile reviewed 2026-09-21 Why consider Jan Desktop Use an open-source desktop client with local and cloud provider choice. Serve a local model to another tool through an OpenAI-compatible endpoint. What to weigh up Using a cloud provider changes the data boundary. Available model size and speed depend on hardware. Models & access Downloads local models and connects to cloud providers; exposes a local OpenAI-compatible server. Local operation can keep inference on-device. Open-source client; inference separate Local hardware or chosen cloud provider supplies inference. [Check current plans and usage terms ↗](https://github.com/janhq/jan) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://www.jan.ai/docs/desktop) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://raw.githubusercontent.com/janhq/jan/main/README.md) Web interface Unknown Open source [Open source ↗](https://raw.githubusercontent.com/janhq/jan/main/README.md) Scope: desktop Multiple model providers [Multiple providers ↗](https://raw.githubusercontent.com/janhq/jan/main/README.md) The desktop client explicitly connects to external providers including OpenAI, Anthropic, Mistral and Groq, separately from downloading local model weights. Cloud requests use the selected external provider. Scope: desktop Local model connection [Documented ↗](https://raw.githubusercontent.com/janhq/jan/main/README.md) Scope: desktop MCP [Documented ↗](https://raw.githubusercontent.com/janhq/jan/main/README.md) Scope: desktop Agent skills [Documented ↗](https://www.jan.ai/docs/agent) Scope: desktop Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://raw.githubusercontent.com/janhq/jan/main/README.md) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Jan official README ↗](https://raw.githubusercontent.com/janhq/jan/main/README.md) official repository · checked 2026-09-21 [Jan Agent and Desktop comparison ↗](https://www.jan.ai/docs/agent) official docs · checked 2026-09-21 | | | | |
| [Jazz](#jazz) AI assistant · lvndry General-purpose local agent for terminal tasks, scripts, scheduled work and messaging integrations. | CLI messaging | Multiple providers Local connection documented | [MIT ↗](https://github.com/lvndry/jazz/blob/main/LICENSE) Open-source client; selected model costs | Compare Jazz Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Jazz Product mark · Profile reviewed 2026-09-21 Why consider Jazz Running one assistant across interactive terminal work and scheduled scripts. Using local models for general assistant tasks. What to weigh up Messaging and email integrations require their own setup. General-purpose assistant scope extends beyond coding; no quality ranking is inferred. Models & access Provider setup includes hosted APIs and local Ollama or llama.cpp. Open-source client; selected model costs Model pricing depends on the chosen provider; local inference requires your own hardware. [Check current plans and usage terms ↗](https://github.com/lvndry/jazz) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/lvndry/jazz) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/lvndry/jazz) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/lvndry/jazz/blob/main/LICENSE) Scope: cli Multiple model providers [Multiple providers ↗](https://github.com/lvndry/jazz) Scope: cli Local model connection [Documented ↗](https://github.com/lvndry/jazz) Scope: cli MCP [Documented ↗](https://github.com/lvndry/jazz) Scope: cli Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Scope: cli Subagents Unknown Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/lvndry/jazz) Skills installation target: .jazz/skills/ · Global: ~/.jazz/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Jazz Readme ↗](https://github.com/lvndry/jazz) official repository · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [MIT license ↗](https://github.com/lvndry/jazz/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Junie](#junie) Coding agent · JetBrains JetBrains coding agent with IDE and terminal workflows and configurable model selection. | IDE CLI | Multiple providers Local connection unknown | Licence not established JetBrains plan or CLI BYOK | Compare Junie Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Junie Product mark · Profile reviewed 2026-09-21 Why consider Junie Use a coding agent within an existing JetBrains workflow. Use the CLI with a provider API key rather than a JetBrains model subscription. What to weigh up The default model is dynamically chosen, not a stable model identifier. Verify IDE BYOK support in the installed plugin; official text currently conflicts. Models & access CLI supports BYOK and custom providers. The plugin models page is internally inconsistent about BYOK, so CLI capabilities must not be advertised as IDE parity. JetBrains plan or CLI BYOK For CLI BYOK, usage is billed by the chosen provider and does not require a JetBrains AI subscription. [Check current plans and usage terms ↗](https://junie.jetbrains.com/docs/byok.html) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://junie.jetbrains.com/docs/junie-cli.html) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://junie.jetbrains.com/docs/junie-cli.html) Editor integration [Documented ↗](https://junie.jetbrains.com/docs/junie-cli.html) Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers [Multiple providers ↗](https://junie.jetbrains.com/docs/byok.html) CLI BYOK/custom provider support verified. Plugin documentation contradicts itself about BYOK. Scope: cli Local model connection Unknown Local plugin support remains unknown because the official plugin page is internally inconsistent. MCP [Documented ↗](https://junie.jetbrains.com/docs/environment-variables.html) CLI environment-variable controls verified; IDE parity not assumed. Scope: cli Agent skills [Documented ↗](https://junie.jetbrains.com/docs/environment-variables.html) CLI skills environment-variable controls verified; IDE parity not assumed. Scope: cli Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://junie.jetbrains.com/docs/junie-cli.html) Skills installation target: .junie/skills/ · Global: ~/.junie/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Junie CLI quickstart ↗](https://junie.jetbrains.com/docs/junie-cli.html) official docs · checked 2026-09-21 [Junie BYOK ↗](https://junie.jetbrains.com/docs/byok.html) official docs · checked 2026-09-21 [Junie plugin models ↗](https://junie.jetbrains.com/docs/junie-plugin-models.html) official docs · checked 2026-09-21 [Junie environment variables ↗](https://junie.jetbrains.com/docs/environment-variables.html) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Kilo Code](#kilo) Coding agent · Kilo / Anaconda Open-source coding platform with editor extensions, a CLI and a model gateway. | IDE CLI | Multiple providers Local connection documented | [MIT ↗](https://github.com/Kilo-Org/kilocode/blob/main/LICENSE) Free client; optional inference, cloud compute and team plans | Compare Kilo Code Add |
| Reasons, limits, setup & sources 7 sources · 2026-09-21 Kilo Code Product mark · Profile reviewed 2026-09-21 Why consider Kilo Code Use a common coding agent through VS Code, JetBrains or terminal. Switch providers/models with a gateway rather than changing the whole client. What to weigh up Model catalogue breadth does not establish task quality. Autonomous command mode can approve actions and requires deliberate permission configuration. Custom subagent documentation specifically covers the CLI; interface parity is not assumed. Models & access Kilo Gateway, direct BYOK, model switching and local Ollama/LM Studio/Atomic Chat are documented. Free client; optional inference, cloud compute and team plans Individuals can use free/local models or BYOK without a paid Kilo inference plan. Gateway usage, optional Kilo Pass credits and cloud compute are separate charges. Credit purchases carry a processing fee. Teams/Enterprise add collaboration and governance. [Check current plans and usage terms ↗](https://kilo.ai/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://kilo.ai/docs) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://raw.githubusercontent.com/Kilo-Org/kilocode/main/README.md) Editor integration [Documented ↗](https://raw.githubusercontent.com/Kilo-Org/kilocode/main/README.md) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://raw.githubusercontent.com/Kilo-Org/kilocode/main/README.md) Public agent code is open source; do not infer hosted gateway infrastructure or every third-party IDE package license. Scope: ide, cli Multiple model providers [Multiple providers ↗](https://raw.githubusercontent.com/Kilo-Org/kilocode/main/README.md) Scope: ide, cli Local model connection [Documented ↗](https://kilo.ai/docs/automate/extending/local-models) Local model settings verified for the IDE; CLI parity is not assumed. Scope: ide MCP [Documented ↗](https://raw.githubusercontent.com/Kilo-Org/kilocode/main/README.md) Scope: ide, cli Agent skills [Documented ↗](https://kilo.ai/docs/customize/skills) Current Skills page explicitly documents CLI and VS Code. Scope: ide, cli Subagents [Documented ↗](https://kilo.ai/docs/customize/custom-subagents) Custom subagent configuration was verified for CLI. Scope: cli Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://raw.githubusercontent.com/Kilo-Org/kilocode/main/README.md) Skills installation target: .agents/skills/ · Global: ~/.kilo/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Kilo official README ↗](https://raw.githubusercontent.com/Kilo-Org/kilocode/main/README.md) official repository · checked 2026-09-21 [Kilo documentation ↗](https://kilo.ai/docs) official docs · checked 2026-09-21 [Kilo skills ↗](https://kilo.ai/docs/customize/skills) official docs · checked 2026-09-21 [Kilo custom subagents ↗](https://kilo.ai/docs/customize/custom-subagents) official docs · checked 2026-09-21 [Kilo local models ↗](https://kilo.ai/docs/automate/extending/local-models) official docs · checked 2026-09-21 [Kilo platform, inference, cloud compute and credit fees ↗](https://kilo.ai/pricing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Kimchi](#kimchi) Coding agent · Kimchi / Cast AI Terminal coding agent using Kimchi-hosted models and optional role-based multi-model orchestration. | CLI | Provider ecosystem Local connection unknown | [Apache-2.0 ↗](https://github.com/getkimchi/kimchi/blob/master/LICENSE) Kimchi service billing | Compare Kimchi Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Kimchi Product mark · Profile reviewed 2026-09-21 Why consider Kimchi Terminal coding with role-based model orchestration. Importing MCP configuration and skills from supported assistants. What to weigh up Requires Kimchi service credentials. Remote-session teleport is explicitly unavailable on Windows in the README. Models & access CLI obtains its supported models from Kimchi metadata service and authenticates with a Kimchi API key. No arbitrary endpoint or local execution support established. Kimchi service billing Kimchi model-service access is separate from the Apache-licensed client. [Check current plans and usage terms ↗](https://github.com/getkimchi/kimchi) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/getkimchi/kimchi) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/getkimchi/kimchi) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/getkimchi/kimchi/blob/master/LICENSE) Multiple model providers [One provider ecosystem ↗](https://github.com/getkimchi/kimchi) Local model connection Unknown MCP [Documented ↗](https://github.com/getkimchi/kimchi) Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Subagents [Documented ↗](https://github.com/getkimchi/kimchi) Hooks / extensions [Documented ↗](https://github.com/getkimchi/kimchi) Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/getkimchi/kimchi) Skills installation target: .kimchi/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Kimchi Readme ↗](https://github.com/getkimchi/kimchi) official repository · checked 2026-09-21 [Kimchi Platform ↗](https://docs.kimchi.dev/docs/kimchi-cli) official docs · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Apache-2.0 license ↗](https://github.com/getkimchi/kimchi/blob/master/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Kimi Code CLI](#kimi-code-cli) Coding agent · Moonshot AI Moonshot’s current standalone terminal coding agent, with plugins, subagents and editor integration through ACP. Also known as Kimi CLI | CLI IDE | Multiple providers Local connection unknown | [MIT ↗](https://github.com/MoonshotAI/kimi-code/blob/main/LICENSE) Kimi membership or provider API usage | Compare Kimi Code CLI Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Kimi Code CLI Product mark · Profile reviewed 2026-09-21 Why consider Kimi Code CLI Use Kimi with native terminal sessions, lifecycle hooks and ACP editors. Migrate existing legacy Kimi CLI data to the supported successor. What to weigh up Windows requires Git Bash. Legacy Python Kimi CLI credentials, MCP authorizations and plugins do not all migrate automatically. Models & access Works with Kimi through OAuth or Moonshot API keys and can be configured for other compatible providers. No unconditional local-model guarantee established. Kimi membership or provider API usage Kimi Code account access uses membership benefits with usage limits. Moonshot Open Platform API access is a separate pay-as-you-go route with different keys and endpoints. Other configured providers follow their own accounts and billing; the CLI software does not include unrestricted hosted inference. [Check current plans and usage terms ↗](https://www.kimi.com/en/help/kimi-code/membership-guide) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/MoonshotAI/kimi-code) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/MoonshotAI/kimi-code) Editor integration [Documented ↗](https://github.com/MoonshotAI/kimi-code) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/MoonshotAI/kimi-code) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Multiple model providers [Multiple providers ↗](https://github.com/MoonshotAI/kimi-code) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Local model connection Unknown The reviewed sources did not establish this capability for a specific interface. MCP [Documented ↗](https://github.com/MoonshotAI/kimi-code) Verified for the terminal agent. Listing an editor/ACP integration does not by itself establish every host editor’s MCP setup or parity. Scope: cli Agent skills [Documented ↗](https://github.com/MoonshotAI/kimi-code) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents [Documented ↗](https://github.com/MoonshotAI/kimi-code) The current Kimi Code repository documents focused subagents with isolated contexts and parallel work. Editor/ACP feature parity was not established. Scope: cli Hooks / extensions [Documented ↗](https://github.com/MoonshotAI/kimi-code) Kimi Code CLI runs local commands at lifecycle events. The scope is the current CLI rather than the archived predecessor or every editor integration. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/MoonshotAI/kimi-code) Skills installation target: .agents/skills/ · Global: ~/.agents/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Current Kimi Code README ↗](https://github.com/MoonshotAI/kimi-code) official repository · checked 2026-09-21 [Legacy archive and migration notice ↗](https://github.com/MoonshotAI/kimi-cli) official repository · checked 2026-09-21 [Kimi Code membership benefits and API billing distinction ↗](https://www.kimi.com/en/help/kimi-code/membership-guide) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Kiro CLI](#kiro-cli) Coding agent · AWS AWS’s terminal coding agent with planning, custom agents, shared Kiro configuration and team controls. Also known as Amazon Q Developer CLI | CLI IDE | Multiple providers Local connection unknown | Licence not established Free tier; paid credit subscriptions | Compare Kiro CLI Add |
| Reasons, limits, setup & sources 7 sources · 2026-09-21 Kiro CLI Product mark · Profile reviewed 2026-09-21 Why consider Kiro CLI Use team steering, custom agents and terminal planning in the Kiro ecosystem. Reuse Kiro configuration while keeping a terminal-focused workflow. What to weigh up Model and preview access varies by plan and region. Administrators requiring a strict model allowlist should check Auto routing semantics. Models & access Kiro’s managed catalog includes multiple vendors with region, plan and governance restrictions. Auto routing can choose generally available regional models outside an administrator’s manually approved list. Free tier; paid credit subscriptions Kiro offers a continuing free tier and paid subscriptions with replenishing credit allocations. Requests consume credits according to the work performed; paid tiers can purchase add-on credits. API-key authentication for the CLI draws from the same subscription credits rather than a separate unrestricted allowance. [Check current plans and usage terms ↗](https://kiro.dev/docs/billing/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://kiro.dev/cli/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://kiro.dev/docs/cli/setup/) Editor integration [Documented ↗](https://kiro.dev/docs/cli/setup/) Desktop app Unknown Web interface Unknown Open source Unknown The reviewed sources did not establish this capability for a specific interface. Multiple model providers [Multiple providers ↗](https://kiro.dev/docs/models/available-models/) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Local model connection Unknown The reviewed sources did not establish this capability for a specific interface. MCP [Documented ↗](https://kiro.dev/docs/cli/setup/) Verified for the terminal agent. Listing an editor/ACP integration does not by itself establish every host editor’s MCP setup or parity. Scope: cli Agent skills [Documented ↗](https://kiro.dev/docs/cli/setup/) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents [Documented ↗](https://kiro.dev/docs/cli/setup/) Kiro CLI setup documents parallel subagents with isolated contexts and a CLI monitoring panel. Native Kiro IDE and ACP integration are distinct surfaces. Scope: cli Hooks / extensions [Documented ↗](https://kiro.dev/docs/hooks/) Kiro documents CLI hooks, with triggers varying by interface. The separate native IDE product's hook support is not transferred to this CLI profile's ACP integration. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://kiro.dev/docs/cli/setup/) Skills installation target: .kiro/skills/ · Global: ~/.kiro/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [CLI setup ↗](https://kiro.dev/docs/cli/setup/) official docs · checked 2026-09-21 [Available models ↗](https://kiro.dev/docs/models/available-models/) official docs · checked 2026-09-21 [Platform requirements ↗](https://kiro.dev/docs/getting-started/installation/) official docs · checked 2026-09-21 [Kiro hooks and interface-specific triggers ↗](https://kiro.dev/docs/hooks/) official docs · checked 2026-09-21 [Kiro billing, tiers and add-on credits ↗](https://kiro.dev/docs/billing/) official docs · checked 2026-09-21 [Kiro CLI authentication and subscription-credit use ↗](https://kiro.dev/docs/cli/authentication/) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Kode](#kode) Coding agent · shareAI-lab Terminal coding agent with configurable models, MCP tools, skills and delegated agent templates. | CLI | Multiple providers Local connection unknown | [Apache-2.0 ↗](https://github.com/shareAI-lab/Kode-Agent/blob/main/LICENSE) Open-source client; model access separate | Compare Kode Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Kode shareAI-lab mark · Profile reviewed 2026-09-21 Why consider Kode Coding with specialized subagent templates and skill/plugin packages. Using explicit safe-mode shell restrictions where supported. What to weigh up Linux system sandbox depends on bubblewrap and mode/configuration. MCP server and provider credentials require configuration. Models & access Model configuration and switching are documented; local weights support was not verified. Open-source client; model access separate Costs depend on the model provider you connect. [Check current plans and usage terms ↗](https://github.com/shareAI-lab/Kode-Agent) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/shareAI-lab/Kode-Agent) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/shareAI-lab/Kode-Agent) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/shareAI-lab/Kode-Agent/blob/main/LICENSE) Multiple model providers [Multiple providers ↗](https://github.com/shareAI-lab/Kode-Agent) Local model connection Unknown MCP [Documented ↗](https://github.com/shareAI-lab/Kode-Agent) Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Subagents [Documented ↗](https://github.com/shareAI-lab/Kode-Agent) Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/shareAI-lab/Kode-Agent) Skills installation target: .kode/skills/ · Global: ~/.kode/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Kode Readme ↗](https://github.com/shareAI-lab/Kode-Agent) official repository · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Apache-2.0 license ↗](https://github.com/shareAI-lab/Kode-Agent/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [LibreChat](#librechat) Model client · LibreChat Self-hosted multi-provider chat and agent platform with file search, skills and subagent support. | web | Multiple providers Local connection documented | [MIT ↗](https://github.com/danny-avila/LibreChat/blob/main/LICENSE) Open-source self-hosting; providers separate | Compare LibreChat Add |
| Reasons, limits, setup & sources 2 sources · 2026-09-21 LibreChat Product mark · Profile reviewed 2026-09-21 Why consider LibreChat Operate a shared model and agent interface under your own administration. Bundle reusable skills and MCP tools into agent workflows. What to weigh up Requires deployment and maintenance of a web application. Attached coding workspaces and some plugin features are explicitly experimental. Models & access Cloud providers and custom OpenAI-compatible endpoints, including local runtimes. Agents can combine files, tools and code execution. Open-source self-hosting; providers separate Hosting and connected model/API services have their own costs. [Check current plans and usage terms ↗](https://github.com/danny-avila/LibreChat) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://www.librechat.ai/docs/quick_start/introduction) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app Unknown Web interface [Documented ↗](https://raw.githubusercontent.com/danny-avila/LibreChat/main/README.md) Open source [Open source ↗](https://raw.githubusercontent.com/danny-avila/LibreChat/main/README.md) Scope: web Multiple model providers [Multiple providers ↗](https://raw.githubusercontent.com/danny-avila/LibreChat/main/README.md) Scope: web Local model connection [Documented ↗](https://raw.githubusercontent.com/danny-avila/LibreChat/main/README.md) Scope: web MCP [Documented ↗](https://raw.githubusercontent.com/danny-avila/LibreChat/main/README.md) Scope: web Agent skills [Documented ↗](https://raw.githubusercontent.com/danny-avila/LibreChat/main/README.md) Scope: web Subagents [Documented ↗](https://raw.githubusercontent.com/danny-avila/LibreChat/main/README.md) Scope: web Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://raw.githubusercontent.com/danny-avila/LibreChat/main/README.md) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [LibreChat official README ↗](https://raw.githubusercontent.com/danny-avila/LibreChat/main/README.md) official repository · checked 2026-09-21 [LibreChat current MIT license ↗](https://raw.githubusercontent.com/danny-avila/LibreChat/main/LICENSE) official repository · checked 2026-09-21 | | | | |
| [llama.cpp](#llama-cpp) Model client · ggml-org C/C++ model inference runtime with terminal chat, an API server and a built-in web chat interface. | CLI web | Local models Local connection documented | [MIT ↗](https://raw.githubusercontent.com/ggml-org/llama.cpp/master/LICENSE) MIT runtime; bring your hardware and model weights | Compare llama.cpp Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 llama.cpp Product mark · Profile reviewed 2026-09-21 Why consider llama.cpp Serve compatible local weights through an API or the bundled browser interface. Tune inference, quantization and CPU/GPU offload directly when runtime control matters. What to weigh up Select compatible model files, hardware backends and memory settings yourself. Tool-call output is a model API capability; it does not by itself provide MCP servers, portable skills or an autonomous coding agent. Models & access Runs compatible local model files across CPU and GPU backends. The server exposes compatible API routes for other clients. MIT runtime; bring your hardware and model weights Software license and individual model-weight licenses are separate. Hosting costs depend on deployment. [Check current plans and usage terms ↗](https://github.com/ggml-org/llama.cpp) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/ggml-org/llama.cpp) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://raw.githubusercontent.com/ggml-org/llama.cpp/master/README.md) Editor integration Unknown Desktop app Unknown Web interface [Documented ↗](https://raw.githubusercontent.com/ggml-org/llama.cpp/master/README.md) Open source [Open source ↗](https://raw.githubusercontent.com/ggml-org/llama.cpp/master/LICENSE) Scope: cli, web Multiple model providers Unknown The CLI and server run compatible model weights; the web chat connects to that inference server. Exposing compatible API routes does not establish integration with multiple external API providers. Local model connection [Documented ↗](https://raw.githubusercontent.com/ggml-org/llama.cpp/master/README.md) Inference occurs where you run llama.cpp; the web interface connects to that server. Downloads and any external integrations need network access. Scope: cli, web MCP Unknown Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current official product documentation verified. [Source ↗](https://github.com/ggml-org/llama.cpp) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [llama.cpp README ↗](https://raw.githubusercontent.com/ggml-org/llama.cpp/master/README.md) official repository · checked 2026-09-21 [llama.cpp server APIs and web UI ↗](https://raw.githubusercontent.com/ggml-org/llama.cpp/master/tools/server/README.md) official repository · checked 2026-09-21 [llama.cpp MIT license ↗](https://raw.githubusercontent.com/ggml-org/llama.cpp/master/LICENSE) official repository · checked 2026-09-21 | | | | |
| [LLM](#llm) AI assistant · Simon Willison CLI and Python library for prompting models, logging conversations, embeddings and repeatable shell/data workflows. | CLI | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/simonw/llm/blob/main/LICENSE) Provider APIs or local-model plugins | Compare LLM Add |
| Reasons, limits, setup & sources 1 sources · 2026-09-21 LLM Neutral lettermark; product logo not established · Profile reviewed 2026-09-21 Why consider LLM Pipe data through model prompts and keep queryable local logs. Build reusable CLI/Python workflows with model and tool plugins. What to weigh up A data/prompting toolkit, not a turnkey autonomous repository agent. Provider and tool capabilities depend on installed plugins. Models & access Hosted providers and local models are added through plugins; Ollama support is explicitly demonstrated. Provider APIs or local-model plugins The CLI uses credentials for the model providers you configure, with provider-specific API charges and limits. Plugins can instead connect to models running on your own device. Installing the client or a provider plugin does not purchase hosted inference. [Check current plans and usage terms ↗](https://github.com/simonw/llm) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://llm.datasette.io/en/stable/setup.html) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/simonw/llm) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/simonw/llm) Multiple model providers [Multiple providers ↗](https://github.com/simonw/llm) Local model connection [Documented ↗](https://github.com/simonw/llm) MCP Unknown Agent skills Unknown Subagents Unknown Hooks / extensions [Documented ↗](https://github.com/simonw/llm) Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/simonw/llm) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [LLM README and documentation index ↗](https://github.com/simonw/llm) official repository · checked 2026-09-21 | | | | |
| [LM Studio](#lm-studio) Model client · Element Labs Local model desktop client and inference server with document chat and developer interfaces. | desktop CLI | Local models Local connection documented | Licence not established Free local app for home/work; optional paid services | Compare LM Studio Add |
| Reasons, limits, setup & sources 6 sources · 2026-09-21 LM Studio Product mark · Profile reviewed 2026-09-21 Why consider LM Studio Download models, compare local behavior and serve them to other applications. Chat with documents offline after downloading the needed models. What to weigh up Desktop licensing is distinct from the licenses of its open-source inference engines and SDKs. Local inference speed and capacity depend on hardware and chosen weights. Models & access Downloaded models run locally; local server offers OpenAI-compatible APIs. Models must fit the available hardware. The lms CLI and headless llmster are documented. Free local app for home/work; optional paid services The local LM Studio app is explicitly free for personal and work use. Paid-feature terms cover optional subscriptions/usage credits. The current main pricing page describes the separate Bionic product and cloud allowances; downloaded model licenses and hardware costs remain separate. [Check current plans and usage terms ↗](https://lmstudio.ai/blog/free-for-work) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://lmstudio.ai/docs/app) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://lmstudio.ai/docs/app) Editor integration Unknown Desktop app [Documented ↗](https://lmstudio.ai/docs/app) Web interface Unknown Open source Unknown Multiple model providers Unknown LM Studio documents loading model weights and serving them through compatible APIs. These are local inference and server features; support for choosing multiple external API providers was not established. Local model connection [Documented ↗](https://lmstudio.ai/docs/app) Desktop and lms/llmster serve downloaded weights locally. Model search/download and remote tools need network. Scope: desktop, cli MCP [Documented ↗](https://lmstudio.ai/docs/app) MCP client verified in desktop chat; this does not automatically give lms CLI the same tool workflow. Scope: desktop Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://lmstudio.ai/docs/app) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [LM Studio app documentation ↗](https://lmstudio.ai/docs/app) official docs · checked 2026-09-21 [Offline operation ↗](https://lmstudio.ai/docs/app/offline) official docs · checked 2026-09-21 [Bionic and LM Studio relationship ↗](https://lmstudio.ai/docs/bionic) official docs · checked 2026-09-21 [LM Studio free home and work use ↗](https://lmstudio.ai/blog/free-for-work) official docs · checked 2026-09-21 [LM Studio current application and paid-feature terms ↗](https://lmstudio.ai/app-terms) official docs · checked 2026-09-21 [Separate Bionic local/free and cloud/paid plans ↗](https://lmstudio.ai/pricing) official docs · checked 2026-09-21 | | | | |
| [LM Studio Bionic](#lm-studio-bionic) AI assistant · Element Labs Agent workspace for code, research and documents using local or hosted open models. | desktop | Local models Local connection documented | Licence not established Local use and optional cloud credits | Compare LM Studio Bionic Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 LM Studio Bionic Product mark · Profile reviewed 2026-09-21 Why consider LM Studio Bionic Work on code or documents with local open models in an agent workspace. Use a remote personal device for inference through LM Link. What to weigh up Cloud use requires account/billing and consumes credits. The model still needs the tool and input capabilities required by the task. Models & access Sessions can use on-device models, another device through LM Link, or LM Studio cloud. Local and remote models do not consume cloud credits. Local use and optional cloud credits Credits are consumed for cloud inference; local and LM Link inference do not require cloud billing. [Check current plans and usage terms ↗](https://lmstudio.ai/docs/bionic/accounts-plans-and-billing/credits-and-usage) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://lmstudio.ai/docs/bionic) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://lmstudio.ai/docs/bionic) Web interface Unknown Open source Unknown Multiple model providers Unknown Documented choices are on-device model weights, a linked device through LM Link, and LM Studio Secure Cloud. These execution locations do not establish integration with multiple external API providers. Local model connection [Documented ↗](https://lmstudio.ai/docs/bionic/models) Scope: desktop MCP Unknown Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://lmstudio.ai/docs/bionic) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Bionic overview ↗](https://lmstudio.ai/docs/bionic) official docs · checked 2026-09-21 [Choose cloud, local or remote models ↗](https://lmstudio.ai/docs/bionic/models) official docs · checked 2026-09-21 [Bionic credits and usage ↗](https://lmstudio.ai/docs/bionic/accounts-plans-and-billing/credits-and-usage) official docs · checked 2026-09-21 | | | | |
| [Loaf](#loaf) Framework / platform · levifig Portable skills, workflow profiles and a local CLI layered over existing coding assistants. | CLI | Not established Local connection unknown | [MIT ↗](https://github.com/levifig/loaf/blob/main/LICENSE) Host-dependent | Compare Loaf Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Loaf Neutral lettermark; product logo not established · Profile reviewed 2026-09-21 Why consider Loaf Sharing skills and tracker-based workflows across supported coding assistants. Local continuity and deterministic checks invoked by host hooks. What to weigh up Needs a supported host assistant. Profile and hook support varies by host; main-branch documentation can describe unreleased capabilities. Models & access Loaf uses the selected host assistant; it is not an independent model runtime. Host-dependent Your host assistant and model access have separate costs. [Check current plans and usage terms ↗](https://github.com/levifig/loaf) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/levifig/loaf) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/levifig/loaf) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/levifig/loaf/blob/main/LICENSE) Multiple model providers Unknown Local model connection Unknown MCP Unknown Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Subagents Unknown Hooks / extensions [Documented ↗](https://github.com/levifig/loaf) Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/levifig/loaf) Skills installation target: .agents/skills/ · Global: ~/.agents/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Loaf Readme ↗](https://github.com/levifig/loaf) official repository · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [MIT license ↗](https://github.com/levifig/loaf/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [MCPJam](#mcpjam) Framework / platform · MCPJam MCP testing and evaluation platform with hosted, desktop and locally served inspector interfaces. | web desktop CLI | Not established Local connection unknown | [Apache-2.0 ↗](https://github.com/MCPJam/inspector/blob/main/LICENSE) Local open source; hosted plans | Compare MCPJam Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 MCPJam Product mark · Profile reviewed 2026-09-21 Why consider MCPJam Debugging MCP tools, resources and authorization exchanges. Running client emulation and regression evaluations for MCP servers. What to weigh up Hosted mode is HTTPS-only and lacks local STDIO capabilities. Local versus hosted skills capabilities differ; README currently contains inconsistent hosted-skills descriptions. Models & access Playground can test models across providers; arbitrary/local endpoint support was not established. Local open source; hosted plans Local use and hosted workspaces have different terms; check the current hosted plan for your team. [Check current plans and usage terms ↗](https://www.mcpjam.com/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/MCPJam/inspector) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/MCPJam/inspector) Editor integration Unknown Desktop app [Documented ↗](https://github.com/MCPJam/inspector) Web interface [Documented ↗](https://github.com/MCPJam/inspector) Open source [Open source ↗](https://github.com/MCPJam/inspector/blob/main/LICENSE) Verified source license applies to local inspector source. Hosted web service licensing is not inferred. Scope: cli Multiple model providers Unknown Local model connection Unknown MCP [Documented ↗](https://github.com/MCPJam/inspector) Scope: cli, desktop, web Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Scope: cli, desktop Subagents Unknown Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/MCPJam/inspector) Skills installation target: .mcpjam/skills/ · Global: ~/.mcpjam/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Mcpjam Readme ↗](https://github.com/MCPJam/inspector) official repository · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Apache-2.0 license ↗](https://github.com/MCPJam/inspector/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [MiniMax Code](#minimax-code) Coding agent · MiniMax Terminal coding agent with interactive, headless and ACP interfaces, with MiniMax account or BYOK access. Also known as MCode · mcode | CLI | Multiple providers Local connection unknown | [MIT ↗](https://github.com/MiniMax-AI/minimax-code/blob/main/LICENSE) MiniMax credits or BYOK | Compare MiniMax Code Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 MiniMax Code Product mark · Profile reviewed 2026-09-21 Why consider MiniMax Code Terminal code changes and tests with MiniMax account or a custom model endpoint. Headless tasks or editor access through ACP. What to weigh up One-command installer excludes Alpine/musl Linux. Published source covers terminal/headless/ACP; desktop app source is explicitly excluded. Models & access Official CLI supports custom providers via --base-url and OpenAI-compatible API formats. Local weights are not explicitly established here. MiniMax credits or BYOK The MiniMax Token Plan requires available credits. With your own provider key, model charges come from that provider and MiniMax login is unnecessary. [Check current plans and usage terms ↗](https://github.com/MiniMax-AI/minimax-code) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/MiniMax-AI/minimax-code) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/MiniMax-AI/minimax-code) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/MiniMax-AI/minimax-code/blob/main/LICENSE) Multiple model providers [Multiple providers ↗](https://github.com/MiniMax-AI/minimax-code) Local model connection Unknown MCP Unknown Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Subagents Unknown Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/MiniMax-AI/minimax-code) Skills installation target: .minimax/skills/ · Global: ~/.minimax/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Minimax Code Readme ↗](https://github.com/MiniMax-AI/minimax-code) official repository · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [MIT license ↗](https://github.com/MiniMax-AI/minimax-code/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Minion Code](#minion-code) Coding agent · femto Configurable coding assistant built on the Minion framework with command-line and ACP operation. Also known as MinionCodeAgent | CLI | Multiple providers Local connection unknown | [MIT ↗](https://github.com/femto/minion-code#license) Open-source client; provider access separate | Compare Minion Code Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Minion Code Neutral lettermark; product logo not established · Profile reviewed 2026-09-21 Why consider Minion Code Using a preconfigured tool-capable coding agent built on the Minion framework. Supplying MCP tool configuration to CLI coding tasks. What to weigh up Setup requires Minion dependency and model configuration files. README mislabels ACP as Agent Communication Protocol while describing Zed-compatible Agent Client Protocol usage; protocol terminology is inconsistent. Models & access CLI documents selecting configured models from multiple vendors. Provider setup is delegated to the Minion dependency; local weights not verified. Open-source client; provider access separate Provider charges depend on configured Minion model services. [Check current plans and usage terms ↗](https://github.com/femto/minion-code) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/femto/minion-code#installation) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/femto/minion-code) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/femto/minion-code#license) Multiple model providers [Multiple providers ↗](https://github.com/femto/minion-code) Local model connection Unknown MCP [Documented ↗](https://github.com/femto/minion-code) Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current primary documentation describes this product; installed behavior and maintenance cadence were not tested. [Source ↗](https://github.com/femto/minion-code) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Minion Code Readme ↗](https://github.com/femto/minion-code) official repository · checked 2026-09-21 [Minion Code Acp Discovery ↗](https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json) registry · checked 2026-09-21 [Minion Code License ↗](https://github.com/femto/minion-code#license) official repository · checked 2026-09-21 | | | | |
| [Mistral Vibe](#mistral-vibe) Coding agent · Mistral AI Mistral’s open-source terminal coding agent with configurable tools, models, skills and agent delegation. | CLI IDE | Multiple providers Local connection unknown | [Apache-2.0 ↗](https://github.com/mistralai/mistral-vibe/blob/main/LICENSE) Included plan usage; optional pay-as-you-go | Compare Mistral Vibe Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Mistral Vibe Product mark · Profile reviewed 2026-09-21 Why consider Mistral Vibe Use Mistral’s open-source coding harness with custom tools and providers. Run interactive or programmatic sessions with skills and delegated tasks. What to weigh up Maintainers officially target Unix; Windows works but has weaker support status. Local compatibility must be checked against the chosen provider adapter. Models & access Model and provider settings are configurable in config.toml. Mistral is the primary onboarding path; local model configuration needs endpoint-specific validation. Included plan usage; optional pay-as-you-go Mistral documents free and paid access modes. Included monthly usage is shared across Studio, the API and Vibe Code and is consumed before pay-as-you-go. Continuing beyond the allowance requires eligible pay-as-you-go access to be enabled; otherwise work stops until usage renews. Other providers use their own credentials. [Check current plans and usage terms ↗](https://docs.mistral.ai/vibe/code/cli/api-keys-profiles) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/mistralai/mistral-vibe) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/mistralai/mistral-vibe) Editor integration [Documented ↗](https://github.com/mistralai/mistral-vibe) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/mistralai/mistral-vibe) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Multiple model providers [Multiple providers ↗](https://github.com/mistralai/mistral-vibe) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Local model connection Unknown The reviewed sources did not establish this capability for a specific interface. MCP [Documented ↗](https://github.com/mistralai/mistral-vibe) Verified for the terminal agent. Listing an editor/ACP integration does not by itself establish every host editor’s MCP setup or parity. Scope: cli Agent skills [Documented ↗](https://github.com/mistralai/mistral-vibe) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents [Documented ↗](https://github.com/mistralai/mistral-vibe) Vibe's CLI supports task delegation to separate subagents, including a read-only exploration agent. Editor/ACP feature parity was not established. Scope: cli Hooks / extensions [Documented ↗](https://github.com/mistralai/mistral-vibe) Vibe documents shell lifecycle hooks in project or user hooks.toml configuration; project hooks require trust. Editor integration parity was not established. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/mistralai/mistral-vibe) Skills installation target: .vibe/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Mistral Vibe README ↗](https://github.com/mistralai/mistral-vibe) official repository · checked 2026-09-21 [Mistral Vibe CLI credentials and plan billing ↗](https://docs.mistral.ai/vibe/code/cli/api-keys-profiles) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Mods](#mods) AI assistant · Charmbracelet Command-line model client designed to transform piped input into text, Markdown or structured output. | CLI | Multiple providers Local connection documented | [MIT ↗](https://github.com/charmbracelet/mods/blob/main/LICENSE) Cloud API usage or local backend | Compare Mods Add |
| Reasons, limits, setup & sources 1 sources · 2026-09-21 Mods Charmbracelet mark · Profile reviewed 2026-09-21 Why consider Mods Add model transformations to existing shell pipelines. Reuse conversations, roles and MCP tools from a small terminal client. What to weigh up Designed around command pipelines rather than an autonomous coding workbench. You configure model access and manage provider charges separately. Models & access Supports cloud APIs and local inference through compatible backends including LocalAI. Cloud API usage or local backend Mods sends prompts and pipeline input to the backend you configure. Cloud routes require the corresponding provider account and are subject to its usage charges and limits. A compatible local backend such as LocalAI shifts inference to your own compute resources. [Check current plans and usage terms ↗](https://github.com/charmbracelet/mods) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/charmbracelet/mods) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/charmbracelet/mods) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/charmbracelet/mods) Multiple model providers [Multiple providers ↗](https://github.com/charmbracelet/mods) Local model connection [Documented ↗](https://github.com/charmbracelet/mods) MCP [Documented ↗](https://github.com/charmbracelet/mods) Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/charmbracelet/mods) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Mods README ↗](https://github.com/charmbracelet/mods) official repository · checked 2026-09-21 | | | | |
| [Moxby](#moxby) AI assistant · Moxby Browser extension for chat, live-site modifications and outcome-tracked automated workflows. | browser-extension | Not established Local connection unknown | Licence not established Free extension; paid creation/chat | Compare Moxby Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Moxby Product mark · Profile reviewed 2026-09-21 Why consider Moxby Modifying browser sites and building browser-based workflow tools. Using reusable mods and measured missions within web work. What to weigh up Chat, creation, missions and Desktop Bridge require paid plans. Free installation only enables eligible marketplace mods. Models & access Official page describes ChatGPT/Claude authentication via Desktop Bridge and other models through OpenRouter; arbitrary endpoint support is not established. Free extension; paid creation/chat Eligible marketplace Mods work with the free extension. Chat, creating Mods, Missions and Desktop Bridge require a paid plan. [Check current plans and usage terms ↗](https://moxby.com/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://moxby.com/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP Unknown Agent skills [Documented ↗](https://moxby.com/) Subagents Unknown Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://moxby.com/) Skills installation target: .moxby/skills/ · Global: ~/.moxby/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Moxby Site ↗](https://moxby.com/) official docs · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Msty Studio](#msty) Model client · Msty Desktop and web workspace combining local/cloud chat, documents, reusable skills and tool-assisted work. Also known as Msty | desktop web | Multiple providers Local connection documented | Licence not established Free Desktop tier; paid Aurum and team licenses | Compare Msty Studio Add |
| Reasons, limits, setup & sources 6 sources · 2026-09-21 Msty Studio Product mark · Profile reviewed 2026-09-21 Why consider Msty Studio Manage local model downloads and engines in a desktop GUI, including importing existing model files. Compare models in split chats, reuse knowledge collections and manage SKILL.md workflows. What to weigh up Web access and some advanced features require a paid plan; local models and Toolbox in Web need Desktop or Sidecar. Data is stored per device/browser and does not automatically synchronize. Cloud providers and remote tools still transmit data. Models & access Desktop installs and manages Ollama, MLX and llama.cpp and connects hosted providers. Web requires a Desktop or Sidecar connection for local models and Toolbox. Free Desktop tier; paid Aurum and team licenses Advanced features and Web access depend on paid plans; hosted model API costs are separate. [Check current plans and usage terms ↗](https://msty.ai/pricing/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://msty.ai/products/studio/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://msty.ai/products/studio/) Web interface [Documented ↗](https://msty.ai/products/studio/) Open source Unknown Multiple model providers [Multiple providers ↗](https://docs.msty.ai/studio/getting-started) Scope: desktop, web Local model connection [Documented ↗](https://docs.msty.ai/studio/managing-models/local-models) Desktop runs local engines. Web uses Desktop/Sidecar to reach local inference, rather than running model weights in the browser. Scope: desktop MCP [Documented ↗](https://docs.msty.ai/studio/toolbox/tools) Local stdio and remote Streamable HTTP are supported; direct SSE is not. Web Toolbox requires a Desktop/Sidecar connection. Scope: desktop, web Agent skills [Documented ↗](https://docs.msty.ai/studio/studios/skills-studio) Skills Studio edits SKILL.md packages for Agent Mode; browser-only Agent parity is not verified. Scope: desktop Subagents Unknown Hooks / extensions Unknown Availability: Current official product documentation verified. [Source ↗](https://msty.ai/products/studio/) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Msty Studio product and installers ↗](https://msty.ai/products/studio/) official docs · checked 2026-09-21 [Msty Desktop/Web capabilities and local storage ↗](https://docs.msty.ai/studio/getting-started) official docs · checked 2026-09-21 [Msty local engine management ↗](https://docs.msty.ai/studio/managing-models/local-models) official docs · checked 2026-09-21 [Msty local and remote MCP ↗](https://docs.msty.ai/studio/toolbox/tools) official docs · checked 2026-09-21 [Msty SKILL.md management ↗](https://docs.msty.ai/studio/studios/skills-studio) official docs · checked 2026-09-21 [Msty Studio plan differences ↗](https://msty.ai/pricing/) official docs · checked 2026-09-21 | | | | |
| [Neovate Code](#neovate) Coding agent · NeovateAI Terminal coding agent for changes, reviews and tests with configurable provider/model access. Also known as Neovate | CLI | Multiple providers Local connection unknown | [MIT ↗](https://github.com/neovateai/neovate-code/blob/main/LICENSE) Open-source client; model access separate | Compare Neovate Code Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Neovate Code NeovateAI mark · Profile reviewed 2026-09-21 Why consider Neovate Code Terminal coding with provider and model selection. Reviewing proposed changes and approving tool calls. What to weigh up Provider credentials are required. Advanced features beyond registry skills compatibility were not verified in this pass. Models & access Official README documents provider login/API-key configuration and model selection; local weights not established. Open-source client; model access separate Costs depend on the model provider you connect. [Check current plans and usage terms ↗](https://github.com/neovateai/neovate-code) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/neovateai/neovate-code) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/neovateai/neovate-code) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/neovateai/neovate-code/blob/main/LICENSE) Multiple model providers [Multiple providers ↗](https://github.com/neovateai/neovate-code) Local model connection Unknown MCP Unknown Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Subagents Unknown Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/neovateai/neovate-code) Skills installation target: .neovate/skills/ · Global: ~/.neovate/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Neovate Readme ↗](https://github.com/neovateai/neovate-code) official repository · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [MIT license ↗](https://github.com/neovateai/neovate-code/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Nova (Compass)](#nova) Coding agent · Compass Agentic Platform Terminal-first coding assistant in the Compass ecosystem, with provider choice and specialized agent tasks. | CLI | Multiple providers Local connection documented | [MIT ↗](https://github.com/Compass-Agentic-Platform/nova/blob/main/LICENSE) Local client; optional Compass service | Compare Nova (Compass) Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Nova (Compass) Compass Agentic Platform mark · Profile reviewed 2026-09-21 Why consider Nova (Compass) Terminal coding with configurable models and specialized agent definitions. Optionally connecting sessions to the Compass platform. What to weigh up Compass cloud session synchronization and account features are separate from the local CLI. README includes comparative context/privacy/performance marketing that was not independently substantiated. Models & access README explicitly documents multiple providers and Ollama with model switching. Local client; optional Compass service Model-provider access and optional Compass subscription services have separate billing terms. [Check current plans and usage terms ↗](https://github.com/Compass-Agentic-Platform/nova) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/Compass-Agentic-Platform/nova) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/Compass-Agentic-Platform/nova) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/Compass-Agentic-Platform/nova/blob/main/LICENSE) Multiple model providers [Multiple providers ↗](https://github.com/Compass-Agentic-Platform/nova) Local model connection [Documented ↗](https://github.com/Compass-Agentic-Platform/nova) MCP Unknown Agent skills [Documented ↗](https://github.com/Compass-Agentic-Platform/nova) Subagents [Documented ↗](https://github.com/Compass-Agentic-Platform/nova) Hooks / extensions Unknown Availability: Current primary documentation describes this product; installed behavior and maintenance cadence were not tested. [Source ↗](https://www.compassap.ai/portfolio/nova.html) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Nova Readme ↗](https://github.com/Compass-Agentic-Platform/nova) official repository · checked 2026-09-21 [Nova Acp Discovery ↗](https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json) registry · checked 2026-09-21 [Nova License ↗](https://github.com/Compass-Agentic-Platform/nova/blob/main/LICENSE) official repository · checked 2026-09-21 | | | | |
| [Ollama](#ollama) Model client · Ollama Local model runtime with a terminal interface, desktop app and API; also offers optional cloud inference and integrations with separate agents. | CLI desktop | Local models Local connection documented | [MIT ↗](https://raw.githubusercontent.com/ollama/ollama/main/LICENSE) Open-source local runtime; optional paid cloud usage | Compare Ollama Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Ollama Product mark · Profile reviewed 2026-09-21 Why consider Ollama Run downloaded models through a local API that other chat clients and coding agents can use. Manage model downloads and launch documented agent integrations from the terminal. What to weigh up It supplies model inference; skills, MCP orchestration and coding workflows belong to the connected agent unless separately documented. Local model size and context must fit available hardware; cloud models and web search require network access. Models & access Choose downloaded model weights or Ollama cloud models. Local serving and cloud inference have different data and billing paths. Open-source local runtime; optional paid cloud usage Local execution uses your hardware. Cloud usage credits and subscriptions are separate. [Check current plans and usage terms ↗](https://ollama.com/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/ollama/ollama) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://raw.githubusercontent.com/ollama/ollama/main/README.md) Editor integration Unknown Desktop app [Documented ↗](https://raw.githubusercontent.com/ollama/ollama/main/README.md) Web interface Unknown Open source [Open source ↗](https://raw.githubusercontent.com/ollama/ollama/main/LICENSE) MIT runtime/CLI; do not infer the license of separately installed third-party clients. Scope: cli Multiple model providers Unknown The reviewed runtime and desktop documentation establishes downloaded model weights and Ollama cloud. Serving OpenAI-compatible or Anthropic-compatible APIs does not establish multiple external API-provider integrations. Local model connection [Documented ↗](https://raw.githubusercontent.com/ollama/ollama/main/README.md) Choose a downloaded local model; disable cloud features for local-only operation. Cloud models and web search do not run offline. Scope: cli, desktop MCP Unknown Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current official product documentation verified. [Source ↗](https://ollama.com) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Ollama README ↗](https://raw.githubusercontent.com/ollama/ollama/main/README.md) official repository · checked 2026-09-21 [Ollama MIT license ↗](https://raw.githubusercontent.com/ollama/ollama/main/LICENSE) official repository · checked 2026-09-21 [Ollama quickstart and agent integrations ↗](https://docs.ollama.com/quickstart) official docs · checked 2026-09-21 [Ollama local-only mode and privacy boundary ↗](https://docs.ollama.com/faq) official docs · checked 2026-09-21 [Ollama pricing structure ↗](https://ollama.com/pricing) official docs · checked 2026-09-21 | | | | |
| [Ona](#ona) Framework / platform · Ona Platform for background coding agents in reproducible cloud or private-cloud environments. | web IDE CLI | Not established Local connection unknown | Licence not established Subscription with pooled usage credits; enterprise contracts | Compare Ona Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Ona Product mark · Profile reviewed 2026-09-21 Why consider Ona Run scheduled or event-triggered work in repeatable development containers. Apply organization policies and use customer-VPC runners. What to weigh up Environment and model usage have different boundaries; a private runner does not prove private inference. Requires cloud environment and integration administration. Models & access Agents run in Ona Cloud or customer AWS/GCP infrastructure. This establishes private execution environments, not on-device model inference. Subscription with pooled usage credits; enterprise contracts Core pools credits for model activity and running environments instead of charging per team seat. Add-on credits are available. An existing Codex subscription can cover model usage while Ona still charges compute; customer VPC deployment and advanced controls belong to Enterprise. [Check current plans and usage terms ↗](https://ona.com/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://ona.com/docs/ona/getting-started) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://ona.com/docs/ona/getting-started) Editor integration [Documented ↗](https://ona.com/docs/ona/getting-started) Desktop app Unknown Web interface [Documented ↗](https://ona.com/docs/ona/getting-started) Open source Unknown Multiple model providers Unknown Local model connection Unknown Customer VPC execution is different from on-device model inference. MCP Unknown Agent skills [Documented ↗](https://ona.com/docs/ona/getting-started) Background/cloud agent configuration; attached IDE or CLI is not evidence of separate local execution. Scope: web Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://ona.com/docs/ona/getting-started) Skills installation target: .ona/skills/ · Global: ~/.ona/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Ona overview ↗](https://ona.com/docs/ona/getting-started) official docs · checked 2026-09-21 [Ona Core usage credits, external subscriptions and Enterprise ↗](https://ona.com/pricing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Open WebUI](#open-webui) Model client · Open WebUI Self-hosted chat and agent interface for local and cloud models with multi-user controls. | web | Multiple providers Local connection documented | [Open-WebUI-custom ↗](https://github.com/open-webui/open-webui/blob/main/LICENSE) Self-hosted; optional enterprise | Compare Open WebUI Add |
| Reasons, limits, setup & sources 1 sources · 2026-09-21 Open WebUI Product mark · Profile reviewed 2026-09-21 Why consider Open WebUI Provide one self-hosted interface for several model providers. Configure user/group access, knowledge and tool integrations. What to weigh up Current license has branding conditions; do not label the whole current codebase MIT. Self-hosting adds deployment, authentication and update responsibilities. Models & access Connects Ollama and OpenAI-compatible APIs; local and cloud endpoints can coexist. Offline operation is supported with local dependencies configured. Self-hosted; optional enterprise Infrastructure and model provider costs are separate. [Check current plans and usage terms ↗](https://github.com/open-webui/open-webui) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.openwebui.com/getting-started/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app Unknown Web interface [Documented ↗](https://raw.githubusercontent.com/open-webui/open-webui/main/README.md) Open source Unknown Current code has multiple licenses with an additional branding requirement; OSI status left unknown. Multiple model providers [Multiple providers ↗](https://raw.githubusercontent.com/open-webui/open-webui/main/README.md) Scope: web Local model connection [Documented ↗](https://raw.githubusercontent.com/open-webui/open-webui/main/README.md) Local runtime integration requires configuring the server-side endpoint; hosted deployment is not on-device by default. Scope: web MCP [Documented ↗](https://raw.githubusercontent.com/open-webui/open-webui/main/README.md) Scope: web Agent skills [Documented ↗](https://raw.githubusercontent.com/open-webui/open-webui/main/README.md) Scope: web Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://raw.githubusercontent.com/open-webui/open-webui/main/README.md) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Open WebUI official README ↗](https://raw.githubusercontent.com/open-webui/open-webui/main/README.md) official repository · checked 2026-09-21 | | | | |
| [OpenClaw](#openclaw) AI assistant · OpenClaw Foundation Self-hosted personal or team assistant that connects local devices, messaging channels and interchangeable model harnesses. | CLI desktop web messaging | Multiple providers Local connection documented | [MIT ↗](https://github.com/openclaw/openclaw/blob/main/LICENSE) Self-hosted; no paid OpenClaw tier | Compare OpenClaw Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 OpenClaw Product mark · Profile reviewed 2026-09-21 Why consider OpenClaw Persistent assistants across chat channels and local devices. Self-hosting with configurable local or hosted model connections. What to weigh up Gateway deployment and provider/chat configuration are required. Self-hosting does not mean prompts stay local when cloud models or messaging services are selected. Models & access Hosted and local model providers and interchangeable harness plugins are documented. Self-hosted; no paid OpenClaw tier The OpenClaw foundation offers no paid tier or hosted service. Third-party model and hosting costs remain separate. [Check current plans and usage terms ↗](https://github.com/openclaw/openclaw) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/openclaw/openclaw) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/openclaw/openclaw) Editor integration Unknown Desktop app [Documented ↗](https://github.com/openclaw/openclaw) Web interface [Documented ↗](https://github.com/openclaw/openclaw) Open source [Open source ↗](https://github.com/openclaw/openclaw/blob/main/LICENSE) Scope: cli, web Multiple model providers [Multiple providers ↗](https://github.com/openclaw/openclaw) Scope: cli, web Local model connection [Documented ↗](https://github.com/openclaw/openclaw) Scope: cli, web MCP Unknown Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Scope: cli, web Subagents Unknown Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/openclaw/openclaw) Skills installation target: skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Openclaw Readme ↗](https://github.com/openclaw/openclaw) official repository · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [MIT license ↗](https://github.com/openclaw/openclaw/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [OpenCode](#opencode) Coding agent · Anomaly Open-source coding agent with terminal and desktop interfaces, provider selection and explicit build/plan modes. | CLI desktop | Multiple providers Local connection documented | [MIT ↗](https://github.com/anomalyco/opencode/blob/dev/LICENSE) BYOK/local; optional hosted plans | Compare OpenCode Add |
| Reasons, limits, setup & sources 8 sources · 2026-09-21 OpenCode Product mark · Profile reviewed 2026-09-21 Why consider OpenCode Choose interactive build or read-only planning agents. Use a terminal client or the documented desktop beta. Choose hosted or local models with native skills and MCP. What to weigh up Desktop is explicitly labeled beta in the fetched README. Provider-specific configuration is required before inference. Models & access Supports hosted providers, custom base URLs, Ollama, LM Studio and llama.cpp. Optional OpenCode Zen/Go services are separate from BYOK and local operation. BYOK/local; optional hosted plans OpenCode Zen and Go are optional paid provider paths; BYOK provider or local compute costs are separate. [Check current plans and usage terms ↗](https://opencode.ai/docs/providers/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/anomalyco/opencode) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/anomalyco/opencode) Editor integration Unknown Desktop app [Documented ↗](https://github.com/anomalyco/opencode) Web interface Unknown Open source [Open source ↗](https://github.com/anomalyco/opencode/blob/dev/LICENSE) The official overview describes the OpenCode terminal and desktop application as open source; repository licence is MIT. This does not grant hosted inference entitlements. Scope: cli, desktop Multiple model providers [Multiple providers ↗](https://github.com/anomalyco/opencode) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Local model connection [Documented ↗](https://github.com/anomalyco/opencode) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli MCP [Documented ↗](https://opencode.ai/docs/mcp-servers/) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Agent skills [Documented ↗](https://opencode.ai/docs/skills/) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents [Documented ↗](https://github.com/anomalyco/opencode) The terminal agent README documents a general subagent for multi-step tasks. Desktop feature parity was not established by this source. Scope: cli Hooks / extensions [Documented ↗](https://opencode.ai/docs/plugins/) OpenCode exposes plugin event hooks in JavaScript/TypeScript. This is not a claim of a shared shell-hook configuration format or independently verified desktop parity. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/anomalyco/opencode) Skills installation target: .agents/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [OpenCode README ↗](https://github.com/anomalyco/opencode) official repository · checked 2026-09-21 [Providers ↗](https://opencode.ai/docs/providers/) official docs · checked 2026-09-21 [Agent Skills ↗](https://opencode.ai/docs/skills/) official docs · checked 2026-09-21 [MCP servers ↗](https://opencode.ai/docs/mcp-servers/) official docs · checked 2026-09-21 [Plugin hooks ↗](https://opencode.ai/docs/plugins/) official docs · checked 2026-09-21 [MIT license ↗](https://github.com/anomalyco/opencode/blob/dev/LICENSE) official repository · checked 2026-09-21 [OpenCode interfaces and client licensing ↗](https://opencode.ai/docs/) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [OpenHands Agent Canvas](#openhands-agent-canvas) Framework / platform · OpenHands Self-hosted browser workspace for coding agents, local or remote execution backends, and recurring engineering automations. | web | Multiple providers Local connection unknown | [MIT ↗](https://github.com/OpenHands/OpenHands/blob/main/LICENSE) Self-hosted or optional cloud | Compare OpenHands Agent Canvas Add |
| Reasons, limits, setup & sources 2 sources · 2026-09-21 OpenHands Agent Canvas Product mark · Profile reviewed 2026-09-21 Why consider OpenHands Agent Canvas Coordinate different coding agents and execution backends from a browser. Schedule engineering automations on a machine or server you control. What to weigh up The project is labeled beta. The simplest unsandboxed setup gives the agent server access to its host; Docker setup is a separate documented option. Models & access Supports OpenHands model configuration and external ACP agents including Claude Code, Codex and Gemini. Model access follows the chosen backend or agent. Self-hosted or optional cloud Model/provider and compute costs are separate from the MIT-licensed application. Optional OpenHands cloud/enterprise backends have their own terms. [Check current plans and usage terms ↗](https://github.com/OpenHands/OpenHands) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/OpenHands/OpenHands) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app Unknown Web interface [Documented ↗](https://github.com/OpenHands/OpenHands) Open source [Open source ↗](https://github.com/OpenHands/OpenHands/blob/main/LICENSE) Multiple model providers [Multiple providers ↗](https://github.com/OpenHands/OpenHands) Local model connection Unknown MCP Unknown Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Official current README labels Agent Canvas beta and documents installation and self-hosting. [Source ↗](https://github.com/OpenHands/OpenHands) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Agent Canvas README ↗](https://github.com/OpenHands/OpenHands) official repository · checked 2026-09-21 [MIT license ↗](https://github.com/OpenHands/OpenHands/blob/main/LICENSE) official repository · checked 2026-09-21 | | | | |
| [OpenHands CLI (legacy)](#openhands) unmaintained Coding agent · OpenHands Legacy terminal and headless interface to OpenHands; maintainers now direct users to Agent Canvas. | CLI IDE web | Multiple providers Local connection unknown | Licence not established Legacy CLI; external model usage | Compare OpenHands CLI (legacy) Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 OpenHands CLI (legacy) Product mark · Profile reviewed 2026-09-21 Why consider OpenHands CLI (legacy) Understand or migrate an existing OpenHands CLI installation. What to weigh up The official repository explicitly says this CLI is no longer actively maintained. Agent Canvas is the suggested successor and has a different product/interface scope. Models & access Legacy CLI configuration supports model IDs and base URLs. Treat remaining docs as migration/reference material, not a supported new-install recommendation. Legacy CLI; external model usage The legacy CLI connects to a configured model provider, whose API access and usage terms still apply. Local or hosted runtime resources are separate costs. The official repository is no longer actively maintained and directs new users to Agent Canvas, so this is a migration reference rather than a current service offer. [Check current plans and usage terms ↗](https://github.com/OpenHands/OpenHands-CLI) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/OpenHands/OpenHands-CLI) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/OpenHands/OpenHands-CLI) Editor integration [Documented ↗](https://github.com/OpenHands/OpenHands-CLI) Desktop app Unknown Web interface [Documented ↗](https://github.com/OpenHands/OpenHands-CLI) Open source Unknown The reviewed sources did not establish this capability for a specific interface. Multiple model providers [Multiple providers ↗](https://github.com/OpenHands/OpenHands-CLI) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Local model connection Unknown The reviewed sources did not establish this capability for a specific interface. MCP [Documented ↗](https://github.com/OpenHands/OpenHands-CLI) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Agent skills Unknown The reviewed sources did not establish this capability for a specific interface. Subagents Unknown Hooks / extensions Unknown Availability: OpenHands-CLI README: project no longer actively maintained; recommends Agent Canvas at OpenHands/OpenHands. Do not keep the CLI installer under a silently repurposed Canvas row. [Source ↗](https://github.com/OpenHands/OpenHands-CLI) Skills installation target: .openhands/skills/ · Global: ~/.openhands/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [OpenHands CLI maintenance notice ↗](https://github.com/OpenHands/OpenHands-CLI) official repository · checked 2026-09-21 [Agent Canvas successor ↗](https://github.com/OpenHands/OpenHands) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Pi](#pi) Coding agent · Earendil Minimal open-source terminal agent designed to be customized through extensions, skills and packages. Also known as pi-mono | CLI | Multiple providers Local connection documented | [MIT ↗](https://github.com/earendil-works/pi/blob/main/LICENSE) Provider accounts, API usage or local compute | Compare Pi Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Pi Product mark · Profile reviewed 2026-09-21 Why consider Pi Build a tailored agent harness instead of accepting a fixed workflow. Use branching sessions, SDK/RPC integration and custom lifecycle extensions. What to weigh up MCP and subagents are not built into the core; install or build extensions for them. Permission gates and plan mode also require a chosen extension or surrounding environment. Models & access Supports multiple hosted providers and Ollama; custom providers can be added through model configuration or extensions. Provider accounts, API usage or local compute Use provider API keys, supported account sign-in or locally served models; costs follow that provider or your hardware. A key exception: Pi's docs say Claude Pro/Max sign-in in third-party harnesses uses per-token extra usage, not the Claude plan's included limits. [Check current plans and usage terms ↗](https://pi.dev/docs/latest/providers) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://pi.dev/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://pi.dev/) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/earendil-works/pi) Multiple model providers [Multiple providers ↗](https://pi.dev/) Local model connection [Documented ↗](https://pi.dev/) MCP [No native MCP ↗](https://pi.dev/) No native MCP; extensions can add it. Agent skills [Documented ↗](https://pi.dev/) Subagents [Not available ↗](https://pi.dev/) No native subagents; extensions can add them. Hooks / extensions [Documented ↗](https://pi.dev/) Hooks are extension lifecycle events, not a claim of another agent’s hook-file compatibility. Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://pi.dev/) Skills installation target: .pi/skills/ · Global: ~/.pi/agent/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Pi product and design choices ↗](https://pi.dev/) official docs · checked 2026-09-21 [Pi repository ↗](https://github.com/earendil-works/pi) official repository · checked 2026-09-21 [Pi provider authentication and billing qualifications ↗](https://pi.dev/docs/latest/providers) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Pochi](#pochi) Coding agent · TabbyML Code agent inside VS Code-compatible editors, with optional CLI tasks and BYOK model access. | IDE CLI | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/TabbyML/pochi/blob/main/LICENSE) BYOK with optional cloud services | Compare Pochi Add |
| Reasons, limits, setup & sources 7 sources · 2026-09-21 Pochi Product mark · Profile reviewed 2026-09-21 Why consider Pochi IDE-native agent tasks with isolated Git worktrees. Using self-hosted models without a Pochi account. What to weigh up Optional collaboration and billing services use an account. Parallel worktree tasks and subagent delegation are separate capabilities; both need clear UI descriptions. Models & access README explicitly allows supported BYOK providers, fine-tuned and self-hosted models; no account required for BYOK. BYOK with optional cloud services Provider costs and optional account-based collaboration/billing are separate. [Check current plans and usage terms ↗](https://github.com/TabbyML/pochi) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.getpochi.com/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/TabbyML/pochi) Editor integration [Documented ↗](https://github.com/TabbyML/pochi) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/TabbyML/pochi/blob/main/LICENSE) Scope: ide Multiple model providers [Multiple providers ↗](https://github.com/TabbyML/pochi) Scope: ide Local model connection [Documented ↗](https://github.com/TabbyML/pochi) Scope: ide MCP [Documented ↗](https://docs.getpochi.com/developer-updates/) Scope: ide Agent skills [Documented ↗](https://docs.getpochi.com/skills/) Scope: ide Subagents [Documented ↗](https://docs.getpochi.com/custom-agent/) Pochi custom-agent definitions and the newTask tool establish delegation in the IDE workflow; CLI parity was not independently established. Scope: ide Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/TabbyML/pochi) Skills installation target: .pochi/skills/ · Global: ~/.pochi/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Pochi Readme ↗](https://github.com/TabbyML/pochi) official repository · checked 2026-09-21 [Pochi Custom Agent ↗](https://docs.getpochi.com/custom-agent/) official docs · checked 2026-09-21 [Pochi Updates ↗](https://docs.getpochi.com/developer-updates/) official docs · checked 2026-09-21 [Pochi Skills ↗](https://docs.getpochi.com/skills/) official docs · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Apache-2.0 license ↗](https://github.com/TabbyML/pochi/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [pool (Poolside)](#poolside) Coding agent · Poolside Terminal coding agent that can also serve or host ACP sessions, with headless execution and configurable model endpoints. Also known as Poolside · pool | CLI | Multiple providers Local connection documented | [Proprietary ↗](https://github.com/poolsideai/pool/blob/main/LICENSE.md) Poolside access or selected provider | Compare pool (Poolside) Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 pool (Poolside) Product mark · Profile reviewed 2026-09-21 Why consider pool (Poolside) Terminal coding with plan/build modes, subagents and custom model endpoints. Using one shell as ACP agent or client for other agent servers. What to weigh up License is all rights reserved under Poolside EULA. Windows support is preview; features change when connected to a different ACP server. Hooks fail open and are explicitly not a security boundary. Models & access README documents OpenRouter, Ollama and arbitrary OpenAI-compatible APIs, including local llama.cpp/vLLM. Poolside access or selected provider Poolside access and any connected model provider have their own service terms and charges. [Check current plans and usage terms ↗](https://poolside.ai/eula) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/poolsideai/pool) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/poolsideai/pool) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Not open source ↗](https://github.com/poolsideai/pool/blob/main/LICENSE.md) Multiple model providers [Multiple providers ↗](https://github.com/poolsideai/pool) Local model connection [Documented ↗](https://github.com/poolsideai/pool) MCP [Documented ↗](https://github.com/poolsideai/pool) Agent skills [Documented ↗](https://github.com/poolsideai/pool) Subagents [Documented ↗](https://github.com/poolsideai/pool) Hooks / extensions [Documented ↗](https://github.com/poolsideai/pool) Availability: Current primary documentation describes this product; installed behavior and maintenance cadence were not tested. [Source ↗](https://poolside.ai) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Poolside Readme ↗](https://github.com/poolsideai/pool) official repository · checked 2026-09-21 [Poolside Acp Discovery ↗](https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json) registry · checked 2026-09-21 [Poolside License ↗](https://github.com/poolsideai/pool/blob/main/LICENSE.md) official repository · checked 2026-09-21 | | | | |
| [Posit Assistant](#posit-assistant) AI assistant · Posit Data-analysis assistant connected to live R and Python sessions, with IDE and terminal interfaces. Also known as Positron Assistant · Databot | IDE TUI CLI desktop web | Multiple providers Local connection documented | Licence not established Managed model pass or provider access | Compare Posit Assistant Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Posit Assistant Posit mark · Profile reviewed 2026-09-21 Why consider Posit Assistant Inspect a live R/Python session during data analysis. Use built-in R, Python, Quarto and Shiny skills. What to weigh up Provider configuration differs across Positron and other interfaces. Do not conflate the current Assistant with the superseded Positron Assistant or Databot. Models & access Posit AI Pass, direct cloud providers and local Ollama are documented; provider availability varies by interface. Managed model pass or provider access Posit AI Pass is an option; direct provider usage or local inference is also documented. [Check current plans and usage terms ↗](https://assistant.posit.co/docs/getting-started/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://assistant.posit.co/docs/getting-started/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://assistant.posit.co/docs/getting-started/) Editor integration [Documented ↗](https://assistant.posit.co/docs/getting-started/) Desktop app [Documented ↗](https://assistant.posit.co/docs/getting-started/) Web interface [Documented ↗](https://assistant.posit.co/docs/getting-started/) Open source Unknown Multiple model providers [Multiple providers ↗](https://assistant.posit.co/docs/getting-started/) Core IDE and terminal configuration supports multiple providers. Separate desktop/web adapters were not independently tested. Scope: ide, cli, tui Local model connection [Documented ↗](https://assistant.posit.co/docs/getting-started/) Ollama is documented in core IDE/terminal configuration; do not infer every hosted interface can access the visitor’s localhost. Scope: ide, cli, tui MCP [Documented ↗](https://assistant.posit.co/docs/reference/mcp-servers/) Scope: ide, cli, tui Agent skills [Documented ↗](https://assistant.posit.co/docs/features/skills/) Scope: ide, cli, tui Subagents Unknown Hooks / extensions [Documented ↗](https://assistant.posit.co/docs/reference/config-file/) Scope: ide, cli, tui Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://assistant.posit.co/docs/getting-started/) Skills installation target: .posit/assistant/skills/ · Global: ~/.posit/assistant/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Posit Assistant getting started ↗](https://assistant.posit.co/docs/getting-started/) official docs · checked 2026-09-21 [Posit MCP servers ↗](https://assistant.posit.co/docs/reference/mcp-servers/) official docs · checked 2026-09-21 [Posit skills ↗](https://assistant.posit.co/docs/features/skills/) official docs · checked 2026-09-21 [Posit configuration ↗](https://assistant.posit.co/docs/reference/config-file/) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [PromptScript](#promptscript) Framework / platform · PromptScript contributors Configuration compiler for sharing instructions, skills and other agent configuration across host coding platforms. | CLI | Not established Local connection unknown | [MIT ↗](https://github.com/mrwogu/promptscript/blob/main/LICENSE) Open-source configuration tool | Compare PromptScript Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 PromptScript Product mark · Profile reviewed 2026-09-21 Why consider PromptScript Keeping instructions and skills consistent across different coding assistants. Validating and compiling target-specific configuration from one source. What to weigh up Requires an actual host coding assistant to perform work. Compiling hooks/MCP/subagent configuration does not make PromptScript an MCP client or subagent runtime. Models & access Configuration compiler, not an inference runtime; model support belongs to generated target configuration. Open-source configuration tool Host assistant and model costs remain separate. [Check current plans and usage terms ↗](https://github.com/mrwogu/promptscript) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/mrwogu/promptscript) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/mrwogu/promptscript) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/mrwogu/promptscript/blob/main/LICENSE) Multiple model providers Unknown Local model connection Unknown MCP Unknown Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Subagents Unknown Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/mrwogu/promptscript) Skills installation target: .agents/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Promptscript Readme ↗](https://github.com/mrwogu/promptscript) official repository · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [MIT license ↗](https://github.com/mrwogu/promptscript/blob/main/LICENSE) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Qoder](#qoder) Coding agent · Bright Zenith Coding platform with terminal and visual clients; CLI supports scheduled work, extensibility and managed model access. | CLI IDE desktop web | Multiple providers Local connection unknown | Licence not established Plan credits; eligible custom providers | Compare Qoder Add |
| Reasons, limits, setup & sources 6 sources · 2026-09-21 Qoder Product mark · Profile reviewed 2026-09-21 Why consider Qoder Use terminal automation, persistent project context and remote/cloud continuation. Configure skills, hooks, MCP and specialist agents. What to weigh up Custom model choices are account/catalog constrained. Windows ARM support is excluded in the current CLI install guide. Models & access CLI supports BYOK only through the live Custom wizard and eligible account plans. The wizard’s catalog controls available providers, models and credential fields; arbitrary local endpoints are not established. Plan credits; eligible custom providers Built-in CLI usage consumes the account's plan credits, with add-on credits tracked separately. Eligible individual accounts can connect supported third-party subscriptions through the BYOK wizard. The live account catalog controls that access; custom-provider credentials do not imply unrestricted model choice or included API usage. [Check current plans and usage terms ↗](https://docs.qoder.com/cli/usage) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.qoder.com/cli/installation) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://docs.qoder.com/cli/overview) Editor integration [Documented ↗](https://docs.qoder.com/cli/overview) Desktop app [Documented ↗](https://docs.qoder.com/cli/overview) Web interface [Documented ↗](https://docs.qoder.com/cli/overview) Open source Unknown The reviewed sources did not establish this capability for a specific interface. Multiple model providers [Multiple providers ↗](https://docs.qoder.com/cli/custom-models) The model-access evidence in this profile covers the terminal client. Other listed product surfaces must be checked separately. CLI supports BYOK only through the live Custom wizard and eligible account plans. The wizard’s catalog controls available providers, models and credential fields; arbitrary local endpoints are not established. Scope: cli Local model connection Unknown The reviewed sources did not establish this capability for a specific interface. MCP [Documented ↗](https://docs.qoder.com/cli/overview) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Agent skills [Documented ↗](https://docs.qoder.com/cli/overview) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents [Documented ↗](https://docs.qoder.com/cli/overview) The international CLI overview explicitly lists subagents as a CLI extension mechanism. Other Qoder interfaces are not covered by this capability scope. Scope: cli Hooks / extensions [Documented ↗](https://docs.qoder.com/cli/overview) The international CLI overview explicitly lists hooks. Desktop, editor and hosted feature parity was not established by that CLI source. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://docs.qoder.com/cli/overview) Skills installation target: .qoder/skills/ · Global: ~/.qoder/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [CLI overview ↗](https://docs.qoder.com/cli/overview) official docs · checked 2026-09-21 [Custom models ↗](https://docs.qoder.com/cli/custom-models) official docs · checked 2026-09-21 [Install and upgrade ↗](https://docs.qoder.com/cli/installation) official docs · checked 2026-09-21 [Qoder product surfaces ↗](https://qoder.com/) official docs · checked 2026-09-21 [Qoder CLI plan and add-on credit usage ↗](https://docs.qoder.com/cli/usage) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Qoder CN](#qoder-cn) renamed Coding agent · Qoder China-region Qoder offering with a documented CLI, automation modes and agent extensibility. Also known as Lingma · Tongyi Lingma | CLI IDE | Not established Local connection unknown | Licence not established Regional subscription with shared credits | Compare Qoder CN Add |
| Reasons, limits, setup & sources 6 sources · 2026-09-21 Qoder CN Product mark · Profile reviewed 2026-09-21 Why consider Qoder CN Use the CN product and documentation for a China-region workflow. What to weigh up Regional offering may have different accounts, models and terms. It should not count as an independent quality competitor to the same product family. Models & access Regional model catalog and authentication must be checked in the CN documentation. Do not import international Qoder BYOK entitlements automatically. Regional subscription with shared credits Qoder CN offers an introductory trial and paid personal or enterprise subscriptions. Credits are shared across the regional CLI, IDE and QoderWork products under the same account. Individual paid features require the appropriate Alibaba Cloud main account; international Qoder entitlements should not be assumed to carry over. [Check current plans and usage terms ↗](https://docs.qoder.cn/en/product-overview/account-and-subscription) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.qoder.cn/cli/installation) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://docs.qoder.cn/cli/overview) Editor integration [Documented ↗](https://docs.qoder.cn/cli/overview) Desktop app Unknown Web interface Unknown Open source Unknown The reviewed sources did not establish this capability for a specific interface. Multiple model providers Unknown The reviewed sources did not establish this capability for a specific interface. Local model connection Unknown The reviewed sources did not establish this capability for a specific interface. MCP [Documented ↗](https://docs.qoder.cn/cli/overview) Verified for the terminal agent. Listing an editor/ACP integration does not by itself establish every host editor’s MCP setup or parity. Scope: cli Agent skills [Documented ↗](https://docs.qoder.cn/cli/overview) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Subagents [Documented ↗](https://docs.qoder.cn/cli/overview) The regional Qoder CLI CN overview explicitly lists subagents. This is regional CLI evidence, not an assumption from the international product. Scope: cli Hooks / extensions [Documented ↗](https://docs.qoder.cn/cli/overview) The regional Qoder CLI CN overview explicitly lists hooks. The editor integration has not been independently verified for this capability. Scope: cli Availability: Alibaba says Tongyi Lingma was renamed to Qoder CN; preserve the old search alias. [Source ↗](https://www.alibabacloud.com/help/en/lingma/introduction-of-lingma) Skills installation target: .qoder/skills/ · Global: ~/.qoder-cn/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Qoder CLI CN overview ↗](https://docs.qoder.cn/cli/overview) official docs · checked 2026-09-21 [Qoder CN product ↗](https://qoder.cn/) official docs · checked 2026-09-21 [CN installation guide ↗](https://docs.qoder.cn/cli/installation) official docs · checked 2026-09-21 [Qoder CN subscription and shared suite credits ↗](https://docs.qoder.cn/en/product-overview/account-and-subscription) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Qoder CN introduction and rename ↗](https://www.alibabacloud.com/help/en/lingma/introduction-of-lingma) official docs · checked 2026-09-21 | | | | |
| [Qwen Code](#qwen-code) Coding agent · Qwen / Alibaba Open-source coding agent with terminal, editor, desktop and web surfaces plus multi-provider configuration. | CLI IDE desktop web | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/QwenLM/qwen-code/blob/main/LICENSE) Coding plan, metered API or local compute | Compare Qwen Code Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Qwen Code Product mark · Profile reviewed 2026-09-21 Why consider Qwen Code Use open-source agent customization with multiple model protocols. Run local-model experiments or use Qwen-hosted access from the same harness. What to weigh up Client capabilities do not guarantee that every model can use every tool. Hosted quotas and BYOK inference are separate from the client license. Models & access Supports Qwen, OpenAI, Anthropic and Gemini protocols plus third-party and local Ollama/vLLM endpoints. Authentication and entitlements differ by provider. Coding plan, metered API or local compute Current authentication supports Alibaba Coding Plan subscriptions, usage-based ModelStudio access, third-party providers and custom endpoints. The former Qwen OAuth free tier has been discontinued. Provider charges or local compute remain separate from the open-source client; choose credentials that match the intended billing endpoint. [Check current plans and usage terms ↗](https://qwenlm.github.io/qwen-code-docs/en/users/configuration/auth/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/QwenLM/qwen-code) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/QwenLM/qwen-code) Editor integration [Documented ↗](https://github.com/QwenLM/qwen-code) Desktop app [Documented ↗](https://github.com/QwenLM/qwen-code) Web interface [Documented ↗](https://github.com/QwenLM/qwen-code) Open source [Open source ↗](https://github.com/QwenLM/qwen-code/blob/main/LICENSE) Apache-2.0 covers the Qwen Code repository and its terminal/web-daemon runtime. Desktop distribution and third-party editor licences were not separately audited. Scope: cli, web Multiple model providers [Multiple providers ↗](https://github.com/QwenLM/qwen-code) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli Local model connection [Documented ↗](https://github.com/QwenLM/qwen-code) Established for this profile’s terminal client. Parity on its other listed interfaces was not separately established in this review. Scope: cli MCP [Documented ↗](https://github.com/QwenLM/qwen-code) Verified in the terminal runtime and the experimental qwen serve web/daemon interface. Desktop/editor feature parity was not separately audited. Scope: cli, web Agent skills [Documented ↗](https://github.com/QwenLM/qwen-code) Verified in the terminal runtime and the experimental qwen serve web/daemon interface. Desktop/editor feature parity was not separately audited. Scope: cli, web Subagents [Documented ↗](https://github.com/QwenLM/qwen-code) The CLI repository documents subagents and agent teams. Its broader list of interfaces does not establish the same feature in every editor, desktop or web client. Scope: cli Hooks / extensions [Documented ↗](https://github.com/QwenLM/qwen-code) The CLI repository explicitly documents hooks. This audit does not transfer CLI hook configuration to Qwen's other interfaces. Scope: cli Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/QwenLM/qwen-code) Skills installation target: .qwen/skills/ · Global: ~/.qwen/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Qwen Code README ↗](https://github.com/QwenLM/qwen-code) official repository · checked 2026-09-21 [Apache license ↗](https://github.com/QwenLM/qwen-code/blob/main/LICENSE) official repository · checked 2026-09-21 [Web/daemon runtime capabilities ↗](https://qwenlm.github.io/qwen-code-docs/en/users/qwen-serve/) official docs · checked 2026-09-21 [Qwen Code current authentication and retired OAuth free tier ↗](https://qwenlm.github.io/qwen-code-docs/en/users/configuration/auth/) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Reasonix](#reasonix) Coding agent · esengine Local coding-agent engine available through terminal, desktop, browser and editor integrations. Also known as DeepSeek-Reasonix | CLI desktop web IDE | Multiple providers Local connection unknown | [MIT ↗](https://github.com/esengine/DeepSeek-Reasonix/blob/main-v2/LICENSE) Open-source client; model access separate | Compare Reasonix Add |
| Reasons, limits, setup & sources 6 sources · 2026-09-21 Reasonix Product mark · Profile reviewed 2026-09-21 Why consider Reasonix Using one local engine across terminal and GUI surfaces. Long coding sessions with checkpoints and configurable subagent profiles. What to weigh up Provider/model configuration is required. Repository and release links use multiple historical names; retain provenance instead of inferring a vendor. Models & access Current README describes configurable providers/model setup and runtime provider extensions; no local-weight claim inferred. Open-source client; model access separate Costs depend on the model provider you connect. [Check current plans and usage terms ↗](https://github.com/esengine/DeepSeek-Reasonix) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/esengine/DeepSeek-Reasonix) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/esengine/DeepSeek-Reasonix) Editor integration [Documented ↗](https://github.com/esengine/DeepSeek-Reasonix) Desktop app [Documented ↗](https://github.com/esengine/DeepSeek-Reasonix) Web interface [Documented ↗](https://github.com/esengine/DeepSeek-Reasonix) Open source [Open source ↗](https://github.com/esengine/DeepSeek-Reasonix/blob/main-v2/LICENSE) Scope: cli Multiple model providers [Multiple providers ↗](https://github.com/esengine/DeepSeek-Reasonix) Scope: cli Local model connection Unknown MCP [Documented ↗](https://github.com/esengine/DeepSeek-Reasonix) Scope: cli Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Scope: cli Subagents [Documented ↗](https://github.com/esengine/DeepSeek-Reasonix/blob/main-v2/docs/SUBAGENT_PROFILES.md) Official subagent-profile documentation explicitly shares isolated child-agent profiles across desktop, interactive CLI and headless CLI. Browser and ACP parity are not asserted here. Scope: cli, desktop Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/esengine/DeepSeek-Reasonix) Skills installation target: .reasonix/skills/ · Global: ~/.reasonix/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Reasonix Readme ↗](https://github.com/esengine/DeepSeek-Reasonix) official repository · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [MIT license ↗](https://github.com/esengine/DeepSeek-Reasonix/blob/main-v2/LICENSE) official repository · checked 2026-09-21 [Reasonix registry integration ↗](https://github.com/vercel-labs/skills/pull/1300) registry · checked 2026-09-21 [Reasonix subagent profiles ↗](https://github.com/esengine/DeepSeek-Reasonix/blob/main-v2/docs/SUBAGENT_PROFILES.md) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Replit](#replit) Framework / platform · Replit Hosted agent and development platform that can build, test and publish applications. | web desktop mobile | Not established Local connection unknown | Licence not established Plan allowance plus paid usage | Compare Replit Add |
| Reasons, limits, setup & sources 2 sources · 2026-09-21 Replit Product mark · Profile reviewed 2026-09-21 Why consider Replit Build and deploy a prototype in a managed environment. Use built-in app testing and connected services without configuring a local development toolchain. What to weigh up Hosted infrastructure and agent usage introduce separate consumption and deployment decisions. This is a cloud development platform, not a substitute for offline terminal access. Models & access Managed model selection and Free, Power and Max modes are documented. No arbitrary local model endpoint was confirmed. Plan allowance plus paid usage Free mode has plan-specific allowances; paid modes consume paid usage. Deployment services may add costs. [Check current plans and usage terms ↗](https://docs.replit.com/features/agent/overview) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.replit.com/features/agent/overview) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://docs.replit.com/features/agent/overview) Web interface [Documented ↗](https://docs.replit.com/features/agent/overview) Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP [Documented ↗](https://docs.replit.com/features/agent/overview) Agent web documentation verifies product MCP capabilities; desktop/mobile surface parity is not assumed. Scope: web Agent skills [Documented ↗](https://docs.replit.com/features/agent/overview) Agent web skills are documented; desktop/mobile parity is not assumed. Scope: web Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://docs.replit.com/features/agent/overview) Skills installation target: .agents/skills/ · Global: $XDG_CONFIG_HOME/agents/skills/ (default ~/.config/agents/skills/) . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Replit Agent overview ↗](https://docs.replit.com/features/agent/overview) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Roo Code](#roo) retired Coding agent · Roo Code Discontinued upstream VS Code coding extension retained as a historical and migration reference. | IDE | Multiple providers Local connection unknown | [Apache-2.0 ↗](https://github.com/RooCodeInc/Roo-Code/blob/main/LICENSE) Retired product | Compare Roo Code Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Roo Code Product mark · Profile reviewed 2026-09-21 Why consider Roo Code Identify the status of an existing installation and plan migration. What to weigh up Official docs say the extension shut down on May 15. The maintainers point users toward ZooCode and Cline instead. Models & access Model-agnostic extension required an external provider. Historical compatibility is not current service support. Retired product Do not treat old plans or installation instructions as an active offer. [Check current plans and usage terms ↗](https://roocodeinc.github.io/Roo-Code/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://roocodeinc.github.io/Roo-Code/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration [Documented ↗](https://roocodeinc.github.io/Roo-Code/) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/RooCodeInc/Roo-Code) Scope: ide Multiple model providers [Multiple providers ↗](https://roocodeinc.github.io/Roo-Code/) Scope: ide Local model connection Unknown MCP [Documented ↗](https://roocodeinc.github.io/Roo-Code/) Scope: ide Agent skills Unknown Subagents [Documented ↗](https://roocodeinc.github.io/Roo-Code/) Scope: ide Hooks / extensions Unknown Availability: Official docs say the Roo Code extension shut down; not eligible for default new-user suggestions. [Source ↗](https://roocodeinc.github.io/Roo-Code/) Skills installation target: .roo/skills/ · Global: ~/.roo/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Roo Code official shutdown notice ↗](https://roocodeinc.github.io/Roo-Code/) official docs · checked 2026-09-21 [Roo Code repository ↗](https://github.com/RooCodeInc/Roo-Code) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Rovo Dev CLI](#rovodev) Coding agent · Atlassian Terminal coding agent integrated with Atlassian work items, documentation and external tools. Also known as Rovo Dev | CLI | Not established Local connection unknown | Licence not established Paid Rovo Dev Standard required | Compare Rovo Dev CLI Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Rovo Dev CLI Atlassian mark · Profile reviewed 2026-09-21 Why consider Rovo Dev CLI Coding linked to Jira work items and Confluence context. Delegation to custom subagents with MCP tools. What to weigh up Official docs require a paid Rovo Dev Standard subscription; CLI is unavailable during the Standard trial. Requires Atlassian authentication and appropriate access. Models & access Model switching is documented; arbitrary provider endpoints and local weights not established. Paid Rovo Dev Standard required CLI access requires paid Rovo Dev Standard and is unavailable during the Standard trial. [Check current plans and usage terms ↗](https://support.atlassian.com/rovo/docs/use-rovo-dev-cli/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://support.atlassian.com/rovo/docs/install-and-run-rovo-dev-cli-on-your-device/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://support.atlassian.com/rovo/docs/use-rovo-dev-cli/) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP [Documented ↗](https://support.atlassian.com/rovo/docs/use-rovo-dev-cli/) Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Subagents [Documented ↗](https://support.atlassian.com/rovo/docs/use-subagents-in-rovo-dev-cli/) Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://support.atlassian.com/rovo/docs/use-rovo-dev-cli/) Skills installation target: .rovodev/skills/ · Global: ~/.rovodev/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Rovo Cli ↗](https://support.atlassian.com/rovo/docs/use-rovo-dev-cli/) official docs · checked 2026-09-21 [Rovo Install ↗](https://support.atlassian.com/rovo/docs/install-and-run-rovo-dev-cli-on-your-device/) official docs · checked 2026-09-21 [Rovo Subagents ↗](https://support.atlassian.com/rovo/docs/use-subagents-in-rovo-dev-cli/) official docs · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Sarvam Code](#sarvam-code) unverified Coding agent · Sarvam AI Announced coding assistant with a live skills-registry target; public availability is not established. | Not established | Not established Local connection unknown | Licence not established Not established | Compare Sarvam Code Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Sarvam Code Sarvam AI mark · Profile reviewed 2026-09-21 Why consider Sarvam Code No source-verified recommendation is available for this record yet. What to weigh up Public documentation is still a coming-soon page. No verified installation, license, pricing or runtime capability details. Models & access Model access has not been established from the sources reviewed. Not established Public documentation describes the product as coming soon; check launch terms when access opens. Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Open official site ↗](https://docs.sarvam.ai/api/sarvam-code) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP Unknown Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Subagents Unknown Hooks / extensions Unknown Availability: Official product documentation says coming soon; registry entry alone is insufficient evidence of availability. [Source ↗](https://docs.sarvam.ai/api/sarvam-code) Skills installation target: .agents/skills/ · Global: ~/.agents/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Sarvam Code Coming Soon ↗](https://docs.sarvam.ai/api/sarvam-code) official docs · checked 2026-09-21 [Sarvam Docs ↗](https://docs.sarvam.ai/) official docs · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [ShellGPT](#shell-gpt) AI assistant · ShellGPT contributors CLI assistant for shell command generation, code snippets and conversational terminal queries. | CLI | Multiple providers Local connection documented | [MIT ↗](https://github.com/TheR1D/shell_gpt/blob/main/LICENSE) Provider API usage; optional local setup | Compare ShellGPT Add |
| Reasons, limits, setup & sources 2 sources · 2026-09-21 ShellGPT Neutral lettermark; product logo not established · Profile reviewed 2026-09-21 Why consider ShellGPT Get short shell commands and explanations in the terminal. Use interactive chat or pipe terminal output into a prompt. What to weigh up Maintainers warn local models may not work as expected. README default-model examples are old; do not treat them as current recommendations. Models & access OpenAI is the default documented path; LiteLLM and local Ollama setups are documented with explicit local-model compatibility warnings. Provider API usage; optional local setup The default OpenAI route requires an API key and paid API usage; a chat subscription is not the same entitlement. Alternative provider and local Ollama setups have their own costs and configuration requirements, and the project explicitly qualifies local-model compatibility. [Check current plans and usage terms ↗](https://github.com/TheR1D/shell_gpt) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/TheR1D/shell_gpt) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/TheR1D/shell_gpt) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/TheR1D/shell_gpt/blob/main/LICENSE) Multiple model providers [Multiple providers ↗](https://github.com/TheR1D/shell_gpt) Local model connection [Documented ↗](https://github.com/TheR1D/shell_gpt) MCP Unknown Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current official product documentation or maintained product repository was accessible on 2026-09-21. [Source ↗](https://github.com/TheR1D/shell_gpt) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [ShellGPT README ↗](https://github.com/TheR1D/shell_gpt) official repository · checked 2026-09-21 [MIT license ↗](https://github.com/TheR1D/shell_gpt/blob/main/LICENSE) official repository · checked 2026-09-21 | | | | |
| [siGit Code](#sigit-code) Coding agent · PT Sigit Mitra Bangun; distribution by Splitfire AB Local-model coding agent with a terminal interface and ACP editor integrations; an optional cloud product is separate. | CLI IDE | Local models Local connection documented | [Apache-2.0 ↗](https://github.com/getsigit/sigit/blob/main/LICENSE) Self-run local client; optional cloud | Compare siGit Code Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 siGit Code Product mark · Profile reviewed 2026-09-21 Why consider siGit Code Coding with on-device weights and no cloud model credential. Using a local agent in ACP-compatible editors. What to weigh up Initial run downloads model weights and requires local compute capacity. Windows supports ACP mode but not the terminal UI. Optional hosted siGit Code Cloud has a separate data/billing path. Models & access Local CLI downloads and runs GGUF models. Hosted siGit Code Cloud is a distinct option; local runtime needs no cloud API key. Self-run local client; optional cloud Local operation requires no subscription or API key. Your hardware and optional hosted cloud service have separate costs. [Check current plans and usage terms ↗](https://github.com/getsigit/sigit) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/getsigit/sigit) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/getsigit/sigit) Editor integration [Documented ↗](https://github.com/getsigit/sigit) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/getsigit/sigit/blob/main/LICENSE) Scope: cli Multiple model providers Unknown Local GGUF inference avoids a hosted provider requirement, but multiple external API-provider integrations are not established by the reviewed documentation. No provider capability scope is asserted. Local model connection [Documented ↗](https://github.com/getsigit/sigit) Scope: cli, ide MCP [Documented ↗](https://github.com/getsigit/sigit) Scope: ide Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current primary documentation describes this product; installed behavior and maintenance cadence were not tested. [Source ↗](https://code.sigit.si) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Sigit Code Readme ↗](https://github.com/getsigit/sigit) official repository · checked 2026-09-21 [Sigit Code Acp Discovery ↗](https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json) registry · checked 2026-09-21 [Sigit Code License ↗](https://github.com/getsigit/sigit/blob/main/LICENSE) official repository · checked 2026-09-21 | | | | |
| [Snowflake CoCo](#cortex) renamed Coding agent · Snowflake Snowflake-integrated coding agent delivered through Snowsight, a desktop IDE and a local CLI. Also known as Cortex Code · Cortex Code CLI · CoCo | CLI desktop IDE web | Not established Local connection unknown | Licence not established Snowflake account/service billing | Compare Snowflake CoCo Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Snowflake CoCo Snowflake mark · Profile reviewed 2026-09-21 Why consider Snowflake CoCo Snowflake SQL, data engineering and analytics workflows. Working with local repositories and Snowflake-governed data from one agent. What to weigh up Requires a Snowflake account, role permissions and service access. Model availability depends on the Snowflake role and account configuration. Models & access Uses Cortex models available to the authenticated Snowflake role. A multi-vendor catalogue does not establish arbitrary endpoint support. Snowflake account/service billing Eligibility and billing depend on your Snowflake account and service usage. [Check current plans and usage terms ↗](https://docs.snowflake.com/en/user-guide/cortex-code/cortex-code) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.snowflake.com/en/user-guide/cortex-code/cortex-code-cli) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://docs.snowflake.com/en/user-guide/cortex-code/cortex-code) Editor integration [Documented ↗](https://docs.snowflake.com/en/user-guide/cortex-code/cortex-code) Desktop app [Documented ↗](https://docs.snowflake.com/en/user-guide/cortex-code/cortex-code) Web interface [Documented ↗](https://docs.snowflake.com/en/user-guide/cortex-code/cortex-code) Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP [Documented ↗](https://docs.snowflake.com/en/user-guide/cortex-code/cortex-code) Scope: cli, desktop, ide Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Scope: cli, desktop, ide Subagents [Documented ↗](https://docs.snowflake.com/en/user-guide/cortex-code/cortex-code) The CoCo overview explicitly documents custom subagents for CLI and Desktop IDE; Snowsight parity is not established. Scope: cli, desktop, ide Hooks / extensions [Documented ↗](https://docs.snowflake.com/en/user-guide/cortex-code/cortex-code) The CoCo overview explicitly documents hooks for CLI and Desktop IDE; Snowsight parity is not established. Scope: cli, desktop, ide Availability: Snowflake explicitly labels CoCo as formerly Cortex Code. Existing cortex executable and registry ID remain. [Source ↗](https://docs.snowflake.com/en/user-guide/cortex-code/cortex-code) Skills installation target: .cortex/skills/ · Global: ~/.snowflake/cortex/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Coco Overview ↗](https://docs.snowflake.com/en/user-guide/cortex-code/cortex-code) official docs · checked 2026-09-21 [Coco Cli ↗](https://docs.snowflake.com/en/user-guide/cortex-code/cortex-code-cli) official docs · checked 2026-09-21 [Coco Rebrand ↗](https://www.snowflake.com/en/product/features/cortex/) official docs · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Stakpak](#stakpak) AI assistant · Stakpak Terminal and background operations agent for infrastructure, deployments and incident investigation. | CLI | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/stakpak/agent/blob/main/LICENSE) Managed access or local/BYOK configuration | Compare Stakpak Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Stakpak Product mark · Profile reviewed 2026-09-21 Why consider Stakpak Infrastructure and CI/CD tasks with operations-specific tooling. Running an ongoing background operations agent with explicit policies. What to weigh up Autopilot requires Docker and additional host-service setup. Research subagents are behind an enable-subagents flag. Rulebooks are documented; Agent Skills format compatibility was not established. Models & access README documents a local OpenAI-compatible provider profile, with Ollama and LM Studio examples, as well as managed access. Managed access or local/BYOK configuration Operating costs depend on your model access and deployment configuration. [Check current plans and usage terms ↗](https://github.com/stakpak/agent) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://stakpak.gitbook.io/docs/get-started/installing-stakpak-cli) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/stakpak/agent) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/stakpak/agent/blob/main/LICENSE) Multiple model providers [Multiple providers ↗](https://github.com/stakpak/agent) Local model connection [Documented ↗](https://github.com/stakpak/agent) MCP [Documented ↗](https://github.com/stakpak/agent) Agent skills Unknown Subagents [Documented ↗](https://github.com/stakpak/agent) Hooks / extensions Unknown Availability: Current primary documentation describes this product; installed behavior and maintenance cadence were not tested. [Source ↗](https://stakpak.dev) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Stakpak Readme ↗](https://github.com/stakpak/agent) official repository · checked 2026-09-21 [Stakpak Acp Discovery ↗](https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json) registry · checked 2026-09-21 [Stakpak License ↗](https://github.com/stakpak/agent/blob/main/LICENSE) official repository · checked 2026-09-21 | | | | |
| [Syncfusion Code Studio](#codestudio) Coding editor · Syncfusion AI coding IDE with Syncfusion component integration, custom agents and local chat model connections. Also known as Code Studio | IDE | Multiple providers Local connection documented | Licence not established Free individual access; paid team seats and model credits | Compare Syncfusion Code Studio Add |
| Reasons, limits, setup & sources 7 sources · 2026-09-21 Syncfusion Code Studio Syncfusion mark · Profile reviewed 2026-09-21 Why consider Syncfusion Code Studio Generate code with direct Syncfusion component context. Use custom instructions, reusable skills and tool hooks in the editor. What to weigh up Ollama support is restricted to Ask and Edit modes in the cited guide. Hardware/model suitability and model license remain separate from the IDE. Models & access Local Ollama is documented, but only for Ask and Edit modes in the current guide. Do not present this as full local agent-mode support. Free individual access; paid team seats and model credits Individual access includes the IDE and server for a single user. Paid team tiers add shared credits, premium models and administration; model requests consume usage credits. BYOK/private models are supported. Starter credits expire, and existing Syncfusion licenses have separate seat entitlements. [Check current plans and usage terms ↗](https://www.syncfusion.com/code-studio/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://help.syncfusion.com/code-studio/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration [Documented ↗](https://help.syncfusion.com/code-studio/) Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers [Multiple providers ↗](https://help.syncfusion.com/code-studio/how-to-guides/add-ollama-model) Scope: ide Local model connection [Documented ↗](https://help.syncfusion.com/code-studio/how-to-guides/add-ollama-model) Ollama is limited to Ask and Edit modes in the current guide; full local Agent mode is not verified. Scope: ide MCP [Documented ↗](https://help.syncfusion.com/code-studio/) Scope: ide Agent skills [Documented ↗](https://help.syncfusion.com/code-studio/) Scope: ide Subagents Unknown Hooks / extensions [Documented ↗](https://help.syncfusion.com/code-studio/) Hooks are documented for agent sessions; their operation with the Ollama Ask/Edit connection is unverified. Scope: ide Combinations to check Not established together: Ollama is documented only for Ask/Edit. Skills require on-demand resource access, but the reviewed documentation does not verify Skills with those local-model modes. [Source ↗](https://help.syncfusion.com/code-studio/how-to-guides/add-ollama-model) Not established together: Ollama is documented only for Ask/Edit. MCP exists in Code Studio, but the reviewed documentation does not verify tool use with its local Ollama connection. [Source ↗](https://help.syncfusion.com/code-studio/how-to-guides/add-ollama-model) Not established together: Ollama is documented only for Ask/Edit, while hooks are described for agent sessions. Their combined operation is not established by the reviewed guides. [Source ↗](https://help.syncfusion.com/code-studio/reference/configure-properties/hooks) Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://help.syncfusion.com/code-studio/) Skills installation target: .codestudio/skills/ · Global: ~/.codestudio/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Code Studio overview ↗](https://help.syncfusion.com/code-studio/) official docs · checked 2026-09-21 [Code Studio Ollama guide ↗](https://help.syncfusion.com/code-studio/how-to-guides/add-ollama-model) official docs · checked 2026-09-21 [Code Studio skills ↗](https://help.syncfusion.com/code-studio/customization/skills) official docs · checked 2026-09-21 [Code Studio hooks prerequisites and lifecycle ↗](https://help.syncfusion.com/code-studio/reference/configure-properties/hooks) official docs · checked 2026-09-21 [Code Studio skill loading and MCP configuration ↗](https://help.syncfusion.com/code-studio/customization/usersettings) official docs · checked 2026-09-21 [Code Studio individual, team, credits and BYOK pricing ↗](https://www.syncfusion.com/code-studio/pricing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Tabnine CLI](#tabnine-cli) Coding agent · Tabnine Terminal coding assistant connected to a team Tabnine deployment and enterprise context. | CLI | Not established Local connection unknown | Licence not established Team/deployment access required | Compare Tabnine CLI Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Tabnine CLI Product mark · Profile reviewed 2026-09-21 Why consider Tabnine CLI Teams already using Tabnine Agents and enterprise context. Applying team coaching guidelines within terminal workflows. What to weigh up Requires access to a Tabnine host and Agents enabled for the team. Installation guide has inconsistent Node minimums; use its prerequisites and verify deployment-specific requirements. Models & access Official product page says developers can select among multiple models; arbitrary endpoints/local weights not established. Team/deployment access required Access depends on your team’s Tabnine deployment and Agents entitlement; confirm it with your administrator. [Check current plans and usage terms ↗](https://docs.tabnine.com/main/getting-started/tabnine-cli) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.tabnine.com/main/getting-started/tabnine-cli/getting-started/installation) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://docs.tabnine.com/main/getting-started/tabnine-cli) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP [Documented ↗](https://www.tabnine.com/platform-cli/) Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Subagents Unknown Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://docs.tabnine.com/main/getting-started/tabnine-cli) Skills installation target: .tabnine/agent/skills/ · Global: ~/.tabnine/agent/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Tabnine Cli ↗](https://docs.tabnine.com/main/getting-started/tabnine-cli) official docs · checked 2026-09-21 [Tabnine Install ↗](https://docs.tabnine.com/main/getting-started/tabnine-cli/getting-started/installation) official docs · checked 2026-09-21 [Tabnine Product ↗](https://www.tabnine.com/platform-cli/) official docs · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Terramind Nucleus](#terramind) unverified Coding editor · Terramind AI-oriented coding IDE in the Terramind product family, with Atlas and external agent integrations. Also known as Terramind | IDE desktop | Not established Local connection unknown | Licence not established Terramind subscription options | Compare Terramind Nucleus Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 Terramind Nucleus Terramind mark · Profile reviewed 2026-09-21 Why consider Terramind Nucleus Reviewing code and agent work in a dedicated workspace. Using supported Claude Code or Codex connections within the Terramind interface. What to weigh up Official page simultaneously advertises platform downloads and a Mac waitlist; immediately installable availability needs confirmation. Cloud automations and local execution have different data paths. Models & access Page describes managed multi-provider models plus Claude Code and Codex integrations. Do not infer arbitrary endpoint or local-weight support. Terramind subscription options Check which features and access options your Terramind plan includes before choosing a subscription. [Check current plans and usage terms ↗](https://terramind.com/nucleus) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://terramind.com/nucleus) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration [Documented ↗](https://terramind.com/nucleus) Desktop app [Documented ↗](https://terramind.com/nucleus) Web interface Unknown Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP [Documented ↗](https://terramind.com/nucleus) Official Nucleus page explicitly lists MCP server support under Claude Code SDK integration; parity across other built-in agent backends is not established. Scope: desktop, ide Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Registry compatibility only; no independently audited Nucleus surface parity. Subagents Unknown Hooks / extensions Unknown Availability: Product identity verified, but download/waitlist messaging conflicts; no installer availability claim. [Source ↗](https://terramind.com/nucleus) Skills installation target: .terramind/skills/ · Global: ~/.terramind/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Terramind Nucleus ↗](https://terramind.com/nucleus) official docs · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Tinycloud](#tinycloud) AI assistant · Aviary Inc. / Cloudglue Specialized agent CLI for video/image analysis and derived clips, subtitles or other content. | CLI | Not established Local connection unknown | Licence not established Cloudglue account/API usage | Compare Tinycloud Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Tinycloud Product mark · Profile reviewed 2026-09-21 Why consider Tinycloud Video and image analysis or content repurposing. Giving an existing coding assistant access to video-work commands through skills. What to weigh up Native Windows is unsupported; use WSL2. Public repository is a launcher/distribution and skills repository; the product binary is proprietary. Models & access Cloud features require a Cloudglue API key; planner provider and local-weight support were not established. Cloudglue account/API usage Cloud features require a Cloudglue API key; account and service usage determine cloud costs. [Check current plans and usage terms ↗](https://github.com/cloudglue/tinycloud) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/cloudglue/tinycloud) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/cloudglue/tinycloud) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Not open source ↗](https://github.com/cloudglue/tinycloud) Multiple model providers Unknown Local model connection Unknown MCP Unknown Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Subagents Unknown Hooks / extensions Unknown Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/cloudglue/tinycloud) Skills installation target: .tinycloud/skills/ · Global: ~/.tinycloud/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Tinycloud Readme ↗](https://github.com/cloudglue/tinycloud) official repository · checked 2026-09-21 [Tinycloud registry identity ↗](https://github.com/vercel-labs/skills/pull/1102) registry · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [TraeCode](#trae) Coding editor · TRAE TRAE coding product; the current homepage separates TraeCode from the TraeWork assistant. Also known as Trae | IDE | Not established Local connection unknown | Licence not established Limited free tier; subscriptions with model-usage allowance | Compare TraeCode Add |
| Reasons, limits, setup & sources 3 sources · 2026-09-21 TraeCode Product mark · Profile reviewed 2026-09-21 Why consider TraeCode Start at the official TraeCode download rather than a similarly named repository. What to weigh up Feature and model details remain insufficiently verified in this snapshot. TraeWork and the separate open-source trae-agent project are not interchangeable with TraeCode. Models & access The public homepage did not expose usable provider documentation in this review. Do not infer local inference or model restrictions. Limited free tier; subscriptions with model-usage allowance Free access is limited to Auto mode and capped usage/autocomplete. Paid tiers unlock the broader model catalogue, larger usage allowances and more TraeWork cloud-task concurrency. These are international TRAE plans, not TRAE China entitlements. [Check current plans and usage terms ↗](https://www.trae.ai/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://www.trae.ai/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration [Documented ↗](https://www.trae.ai/) Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP Unknown Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://www.trae.ai/) Skills installation target: .trae/skills/ · Global: ~/.trae/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [TRAE official product homepage ↗](https://www.trae.ai/) official docs · checked 2026-09-21 [TRAE international free and paid plans ↗](https://www.trae.ai/pricing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [TraeCode CN](#trae-cn) Coding editor · TRAE Chinese TRAE coding product with an IDE and separately documented terminal client. Also known as Trae CN | IDE CLI TUI | Not established Local connection unknown | Licence not established Credit-based memberships; enterprise seats plus usage | Compare TraeCode CN Add |
| Reasons, limits, setup & sources 7 sources · 2026-09-21 TraeCode CN Product mark · Profile reviewed 2026-09-21 Why consider TraeCode CN Use Chinese-language official documentation for the regional product. What to weigh up Global and China offerings may differ; do not transfer model or pricing claims between them. CLI feature documentation does not establish identical IDE behavior. Models & access CLI provides a model selector. Arbitrary local providers and regional parity are not established by this page. Credit-based memberships; enterprise seats plus usage Personal memberships provide usage credits and permit top-ups while active. Lite credits are TraeWork-only; TraeCode needs an applicable plan/general credits. Enterprise, including CLI, combines prepaid seat subscriptions with separately enabled metered overage. Regional plans differ from international TRAE. [Check current plans and usage terms ↗](https://docs.trae.cn/ide_plans-and-billing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://www.trae.cn/) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://www.trae.cn/) Editor integration [Documented ↗](https://www.trae.cn/) Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP [Documented ↗](https://docs.trae.cn/cli_slash-commands) Verified in TRAE China CLI/TUI commands, not inherited by the global product. Scope: cli, tui Agent skills [Documented ↗](https://docs.trae.cn/cli_slash-commands) Verified in TRAE China CLI/TUI slash commands; global TRAE and IDE parity are not assumed. Scope: cli, tui Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://www.trae.cn/) Skills installation target: .trae/skills/ · Global: ~/.trae-cn/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [TRAE China official homepage ↗](https://www.trae.cn/) official docs · checked 2026-09-21 [TraeCode CLI commands ↗](https://docs.trae.cn/cli_slash-commands) official docs · checked 2026-09-21 [TraeCode CLI configuration ↗](https://docs.trae.cn/cli_config-file) official docs · checked 2026-09-21 [TRAE China personal credit plans and product eligibility ↗](https://docs.trae.cn/ide_plans-and-billing) official docs · checked 2026-09-21 [TRAE China enterprise seat subscriptions and CLI access ↗](https://docs.trae.cn/enterprise_subscription-rules) official docs · checked 2026-09-21 [TRAE China enterprise metered overage ↗](https://docs.trae.cn/enterprise_pay-as-you-go) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [v0](#v0) Framework / platform · Vercel Vercel app-building agent for interfaces, prototypes and full-stack applications. | web | Not established Local connection unknown | Licence not established Free allowance and paid credit plans | Compare v0 Add |
| Reasons, limits, setup & sources 2 sources · 2026-09-21 v0 Product mark · Profile reviewed 2026-09-21 Why consider v0 Create an interface from a mockup and iterate with a live preview. Move a web prototype into Vercel deployment or a reviewable pull request. What to weigh up Hosted workflow and model credits differ from local development tools. Paid tiers differ in data controls, team access and credit policies. Models & access The product exposes a managed model selector; local inference and arbitrary provider configuration were not verified. Free allowance and paid credit plans Generations draw credits; Premium is being sunsetted for new users according to current docs. [Check current plans and usage terms ↗](https://v0.app/docs/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://v0.app/docs) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app Unknown Web interface [Documented ↗](https://v0.app/docs) Open source Unknown Multiple model providers Unknown Local model connection Unknown MCP Unknown Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://v0.app/docs) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [v0 documentation ↗](https://v0.app/docs) official docs · checked 2026-09-21 [v0 pricing ↗](https://v0.app/docs/pricing) official docs · checked 2026-09-21 | | | | |
| [VT Code](#vt-code) Coding agent · VT Code contributors Rust terminal coding agent with configurable model providers, lifecycle hooks and MCP/ACP integrations. Also known as vtcode | CLI | Multiple providers Local connection documented | [MIT OR Apache-2.0 ↗](https://github.com/vinhnx/VTCode#license) Open-source client; model access separate | Compare VT Code Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 VT Code Product mark · Profile reviewed 2026-09-21 Why consider VT Code Terminal coding with custom or local model services. Adding skills, MCP tools and approved lifecycle hooks to repeatable workflows. What to weigh up Windows release artifacts are best-effort and may lag macOS/Linux. Some automation flows are experimental. README labels some subscription OAuth access an unofficial compatibility mechanism and recommends provider API keys for a supported path. Models & access Official README and installation guide list OpenAI-compatible custom providers and local Ollama, LM Studio and llama.cpp. Open-source client; model access separate Hosted model services and local inference hardware have separate costs. [Check current plans and usage terms ↗](https://github.com/vinhnx/VTCode) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://github.com/vinhnx/VTCode/blob/main/docs/installation/README.md) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/vinhnx/VTCode) Editor integration Unknown Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://github.com/vinhnx/VTCode#license) Multiple model providers [Multiple providers ↗](https://github.com/vinhnx/VTCode) Local model connection [Documented ↗](https://github.com/vinhnx/VTCode) MCP [Documented ↗](https://github.com/vinhnx/VTCode) Agent skills [Documented ↗](https://github.com/vinhnx/VTCode) Subagents Unknown Hooks / extensions [Documented ↗](https://github.com/vinhnx/VTCode) Availability: Current primary documentation describes this product; installed behavior and maintenance cadence were not tested. [Source ↗](https://github.com/vinhnx/VTCode) Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Vtcode Readme ↗](https://github.com/vinhnx/VTCode) official repository · checked 2026-09-21 [Vtcode Install ↗](https://github.com/vinhnx/VTCode/blob/main/docs/installation/README.md) official repository · checked 2026-09-21 [Vt Code Acp Discovery ↗](https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json) registry · checked 2026-09-21 [Vt Code License ↗](https://github.com/vinhnx/VTCode#license) official repository · checked 2026-09-21 | | | | |
| [Warp](#warp) Coding agent · Warp Terminal and development workspace with built-in agents, a standalone CLI and cloud automation. | desktop CLI web | Multiple providers Local connection unknown | [AGPL-3.0 ↗](https://github.com/warpdotdev/warp/blob/master/LICENSE-AGPL) Plans, BYOK and applicable platform usage | Compare Warp Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Warp Product mark · Profile reviewed 2026-09-21 Why consider Warp Run agents against real terminal programs and inspect their output. Use a terminal workspace with review tools, or the agent CLI in another terminal. What to weigh up The agent harness runs on Warp servers, including when using custom inference endpoints. Local models require a publicly reachable HTTPS endpoint; not an offline/private-local choice. Model inference charges and applicable platform charges are separate. Models & access BYOK and OpenAI-compatible custom endpoints are supported, but the harness runs on Warp servers. Local/private URLs are rejected; a local model must be exposed through public HTTPS. This does not provide offline inference. Plans, BYOK and applicable platform usage Endpoint inference is billed by the provider; Business/Enterprise endpoint runs may still use platform credits. [Check current plans and usage terms ↗](https://docs.warp.dev/agents/inference/custom-inference-endpoint/) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://raw.githubusercontent.com/warpdotdev/warp/master/README.md) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://docs.warp.dev/_llms-txt/agents.txt) Editor integration Unknown Desktop app [Documented ↗](https://docs.warp.dev/_llms-txt/agents.txt) Web interface [Documented ↗](https://docs.warp.dev/_llms-txt/agents.txt) Open source [Open source ↗](https://raw.githubusercontent.com/warpdotdev/warp/master/README.md) Desktop client code: mostly AGPL with MIT UI framework crates. Hosted agent backend is not covered by the client license. Scope: desktop Multiple model providers [Multiple providers ↗](https://docs.warp.dev/agents/inference/custom-inference-endpoint/) Desktop and CLI can use hosted models/BYOK; inference requests route through Warp infrastructure. Scope: desktop, cli Local model connection Unknown Custom endpoint requires public HTTPS; localhost/private addresses are rejected. No offline badge. MCP [Documented ↗](https://docs.warp.dev/_llms-txt/agents.txt) Scope: desktop, cli Agent skills [Documented ↗](https://docs.warp.dev/_llms-txt/agents.txt) Scope: desktop, cli Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://raw.githubusercontent.com/warpdotdev/warp/master/README.md) Skills installation target: .agents/skills/ · Global: ~/.agents/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Warp official README ↗](https://raw.githubusercontent.com/warpdotdev/warp/master/README.md) official repository · checked 2026-09-21 [Warp agent documentation ↗](https://docs.warp.dev/_llms-txt/agents.txt) official docs · checked 2026-09-21 [Warp documentation index ↗](https://docs.warp.dev/llms.txt) official docs · checked 2026-09-21 [Warp custom inference endpoints ↗](https://docs.warp.dev/agents/inference/custom-inference-endpoint/) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Xum](#mux) renamed Coding agent · Coder Open-source desktop agent multiplexer with isolated Git workspaces and local or remote execution. Also known as Mux | desktop web | Multiple providers Local connection documented | [AGPL-3.0 ↗](https://github.com/coder/mux/blob/main/LICENSE) Open-source client; model costs separate | Compare Xum Add |
| Reasons, limits, setup & sources 2 sources · 2026-09-21 Xum Product mark · Profile reviewed 2026-09-21 Why consider Xum Manage separate agent tasks while inspecting Git divergence. Run the same workspace model locally or through SSH. What to weigh up Official README download instructions list macOS and Linux; native Windows support is unconfirmed. Running multiple workspaces does not itself guarantee coordinated or correct agents. Models & access Multiple cloud providers, OpenRouter and local Ollama are documented. Work can run locally, in Git worktrees or through SSH. Open-source client; model costs separate Supply provider access or local inference. [Check current plans and usage terms ↗](https://github.com/coder/mux) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://mux.coder.com/install) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://raw.githubusercontent.com/coder/mux/main/README.md) Web interface [Documented ↗](https://raw.githubusercontent.com/coder/mux/main/README.md) Open source [Open source ↗](https://raw.githubusercontent.com/coder/mux/main/README.md) Scope: desktop, web Multiple model providers [Multiple providers ↗](https://raw.githubusercontent.com/coder/mux/main/README.md) Scope: desktop, web Local model connection [Documented ↗](https://raw.githubusercontent.com/coder/mux/main/README.md) Local Ollama with the desktop app; remotely hosted web UI does not prove model inference on the browser device. Scope: desktop MCP Unknown Agent skills Unknown Subagents Unknown Hooks / extensions Unknown Availability: Official README says Mux was renamed to Xum following a trademark concern. [Source ↗](https://raw.githubusercontent.com/coder/mux/main/README.md) Skills installation target: .mux/skills/ · Global: ~/.mux/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Xum official README and rename ↗](https://raw.githubusercontent.com/coder/mux/main/README.md) official repository · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [ZCode](#zcode) Coding agent · Z.ai Agent-oriented coding workspace with desktop, browser and terminal interfaces. | desktop web CLI | Multiple providers Local connection documented | [Apache-2.0 ↗](https://github.com/zai-org/ZCode/blob/main/LICENSE) Model plan or BYOK | Compare ZCode Add |
| Reasons, limits, setup & sources 8 sources · 2026-09-21 ZCode Product mark · Profile reviewed 2026-09-21 Why consider ZCode Agent-led coding across desktop, browser and terminal interfaces. Using GLM plans or custom compatible model connections. What to weigh up Model plan, quota and provider configuration are separate from the client. Protocol compatibility affects model capabilities; model names alone do not guarantee image or reasoning support. Models & access Official model configuration explicitly permits custom Anthropic/OpenAI-compatible services, including self-hosted services on private networks. Model plan or BYOK Coding plans and API access have separate billing terms; check the option used by your selected provider. [Check current plans and usage terms ↗](https://zcode.z.ai/en/docs/configuration) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://zcode.z.ai/en) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI [Documented ↗](https://github.com/zai-org/ZCode) Editor integration Unknown Desktop app [Documented ↗](https://github.com/zai-org/ZCode) Web interface [Documented ↗](https://github.com/zai-org/ZCode) Open source [Open source ↗](https://github.com/zai-org/ZCode/blob/main/LICENSE) Scope: cli, desktop, web Multiple model providers [Multiple providers ↗](https://zcode.z.ai/en/docs/configuration) Scope: desktop, web Local model connection [Documented ↗](https://zcode.z.ai/en/docs/configuration) Scope: desktop, web MCP [Documented ↗](https://zcode.z.ai/en/docs/skill) Scope: desktop, web Agent skills [Documented ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) Scope: desktop, web Subagents [Documented ↗](https://zcode.z.ai/en/docs/subagents) Documented for the ZCode Agent settings interface in desktop/web. Custom user-level subagents are beta; CLI parity is not asserted. Scope: desktop, web Hooks / extensions [Documented ↗](https://zcode.z.ai/en/docs/hooks) Documented for desktop/web ZCode Agent sessions and settings. User and plugin hooks run; project-level hook configuration is currently ignored. Scope: desktop, web Availability: Current primary documentation describes the product; maintenance cadence and installed runtime were not audited. [Source ↗](https://github.com/zai-org/ZCode) Skills installation target: .zcode/skills/ · Global: ~/.zcode/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Zcode Readme ↗](https://github.com/zai-org/ZCode) official repository · checked 2026-09-21 [Zcode Models ↗](https://zcode.z.ai/en/docs/configuration) official docs · checked 2026-09-21 [Zcode Skills ↗](https://zcode.z.ai/en/docs/skill) official docs · checked 2026-09-21 [Vercel skills agent target source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 [Apache-2.0 license ↗](https://github.com/zai-org/ZCode/blob/main/LICENSE) official repository · checked 2026-09-21 [ZCode subagents ↗](https://zcode.z.ai/en/docs/subagents) official docs · checked 2026-09-21 [ZCode hooks ↗](https://zcode.z.ai/en/docs/hooks) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Zed](#zed) Coding editor · Zed Industries Open-source code editor with native agents, external ACP agents and terminal threads. | IDE | Multiple providers Local connection documented | [GPL-3.0-or-later ↗](https://github.com/zed-industries/zed/blob/main/LICENSE-GPL) Free editor and BYOK; optional hosted-AI subscription | Compare Zed Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Zed Product mark · Profile reviewed 2026-09-21 Why consider Zed Choose a native or external agent without leaving the editor. Use local inference, an existing subscription or a provider API key. What to weigh up External agents and the native agent do not share all capabilities or settings. The editor is open source; hosted model use has separate terms and costs. Models & access Native agent supports hosted models, API keys, subscriptions, gateways and local models. External ACP agents retain their own configuration and provider boundaries. Free editor and BYOK; optional hosted-AI subscription Personal supports your API keys/external agents with capped edit predictions. Pro adds unlimited predictions and an included hosted-model allowance; overages are billed separately with a service markup. Business adds organization controls and does not bundle fixed model credits. Provider-key usage is billed by the provider. [Check current plans and usage terms ↗](https://zed.dev/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://zed.dev/download) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration [Documented ↗](https://raw.githubusercontent.com/zed-industries/zed/main/README.md) Desktop app Unknown Web interface Unknown Open source [Open source ↗](https://raw.githubusercontent.com/zed-industries/zed/main/README.md) Scope: ide Multiple model providers [Multiple providers ↗](https://zed.dev/docs/ai/overview) Native agent provider configuration; external ACP agents retain their own model/settings scope. Scope: ide Local model connection [Documented ↗](https://zed.dev/docs/ai/overview) Scope: ide MCP [Documented ↗](https://zed.dev/docs/ai/overview) Scope: ide Agent skills [Documented ↗](https://zed.dev/docs/ai/overview) Scope: ide Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://raw.githubusercontent.com/zed-industries/zed/main/README.md) Skills installation target: .agents/skills/ · Global: ~/.agents/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Zed official README ↗](https://raw.githubusercontent.com/zed-industries/zed/main/README.md) official repository · checked 2026-09-21 [Zed AI overview ↗](https://zed.dev/docs/ai/overview) official docs · checked 2026-09-21 [Zed Personal, Pro and Business cost boundaries ↗](https://zed.dev/pricing) official docs · checked 2026-09-21 [Zed usage invoicing and spend limits ↗](https://zed.dev/docs/ai/billing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Zencoder](#zencoder) Coding agent · Zencoder IDE agents for code generation, repository context and configurable model access. | IDE | Multiple providers Local connection documented | Licence not established Subscription credits or BYOK | Compare Zencoder Add |
| Reasons, limits, setup & sources 4 sources · 2026-09-21 Zencoder Product mark · Profile reviewed 2026-09-21 Why consider Zencoder Add agent workflows to VS Code, JetBrains or Android Studio. Configure local and enterprise endpoints through explicit provider settings. What to weigh up Managed model credits and BYOK have different billing paths. Zenflow desktop orchestration is a related but separately described surface. Models & access BYOM can use local Ollama, private/VPC and external API endpoints. Global and project provider settings are supported. Subscription credits or BYOK Paid plans include credits; top-ups and model multipliers affect usage. Free access is described as BYOK. [Check current plans and usage terms ↗](https://docs.zencoder.ai/faq/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.zencoder.ai/welcome) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration [Documented ↗](https://docs.zencoder.ai/welcome) Desktop app Unknown Web interface Unknown Open source Unknown Multiple model providers [Multiple providers ↗](https://docs.zencoder.ai/features/custom-models-configuration) Scope: ide Local model connection [Documented ↗](https://docs.zencoder.ai/features/custom-models-configuration) Scope: ide MCP [Documented ↗](https://docs.zencoder.ai/welcome) Scope: ide Agent skills [Documented ↗](https://docs.zencoder.ai/welcome) Scope: ide Subagents Unknown Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://docs.zencoder.ai/welcome) Skills installation target: .zencoder/skills/ · Global: ~/.zencoder/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Zencoder product surfaces ↗](https://docs.zencoder.ai/welcome) official docs · checked 2026-09-21 [Custom model configuration ↗](https://docs.zencoder.ai/features/custom-models-configuration) official docs · checked 2026-09-21 [Zencoder pricing ↗](https://docs.zencoder.ai/faq/pricing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |
| [Zenflow](#zenflow) Framework / platform · Zencoder Desktop orchestration for coding and work tasks, with worktrees and configurable multi-model stages. | desktop | Multiple providers Local connection documented | Licence not established Subscription credits or BYOK | Compare Zenflow Add |
| Reasons, limits, setup & sources 5 sources · 2026-09-21 Zenflow Zencoder mark · Profile reviewed 2026-09-21 Why consider Zenflow Separate planning, implementation and review into explicit workflow stages. Run isolated agent tasks in Git worktrees and inspect resulting changes. What to weigh up Requires a Zencoder account and provider/billing configuration. Pipeline stages and model choice add setup; they do not prove better task outcomes. Models & access Uses Zencoder-managed models and configurable providers; local/private endpoints are documented in the shared configuration. Subscription credits or BYOK Shares Zencoder billing; consumption depends on model and task context. [Check current plans and usage terms ↗](https://docs.zencoder.ai/faq/pricing) Set up on your computer Choose the operating system of the computer you want to install on. Browser tools may not need installation. [Official setup instructions ↗](https://docs.zencoder.ai/get-started/install-zenflow) Documented capabilities Support can depend on the interface, model or plan. Read the qualifications above. Unknown means the reviewed sources did not settle the question. Terminal / CLI Unknown Editor integration Unknown Desktop app [Documented ↗](https://docs.zencoder.ai/get-started/install-zenflow) Web interface Unknown Open source Unknown Multiple model providers [Multiple providers ↗](https://docs.zencoder.ai/features/custom-models-configuration) Scope: desktop Local model connection [Documented ↗](https://docs.zencoder.ai/features/custom-models-configuration) Scope: desktop MCP [Documented ↗](https://docs.zencoder.ai/llms.txt) Scope: desktop Agent skills [Documented ↗](https://docs.zencoder.ai/llms.txt) Scope: desktop Subagents [Documented ↗](https://docs.zencoder.ai/llms.txt) Scope: desktop Hooks / extensions Unknown Availability: Current official documentation is available; this is not a guarantee of support or performance. [Source ↗](https://docs.zencoder.ai/get-started/install-zenflow) Skills installation target: .zencoder/skills/ · Global: ~/.zencoder/skills/ . A registry path is compatibility information, not a quality endorsement. Sources & review Profile reviewed 2026-09-21 . First-party documentation establishes product claims; it is not an independent quality test. [Zenflow installation ↗](https://docs.zencoder.ai/get-started/install-zenflow) official docs · checked 2026-09-21 [Zencoder official documentation index ↗](https://docs.zencoder.ai/llms.txt) official docs · checked 2026-09-21 [Custom models ↗](https://docs.zencoder.ai/features/custom-models-configuration) official docs · checked 2026-09-21 [Zencoder pricing ↗](https://docs.zencoder.ai/faq/pricing) official docs · checked 2026-09-21 [Skills CLI compatibility target and paths ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts) registry · checked 2026-09-21 | | | | |

03 / THE REASONING

## A shortlist you can question.

A useful recommendation names the job, states the requirements, and shows what you give up.

### Why were four tools highlighted in green?

The previous page highlighted Cline, Goose, OpenCode and Qwen Code because their recorded features covered other rows. That calculation mixed desktop apps with IDEs and favoured well-documented tools. It did not measure coding success, reliability or cost. We removed that implied endorsement.

### What makes a tool a match now?

Choose a workflow and requirements. A current, verified tool qualifies only when the reviewed sources document every requirement for the same interface. A CLI licence or local connection is not attributed to a desktop app by association. All qualifying tools appear alphabetically. There is no paid placement, arbitrary top-four cutoff or hidden overall score.

### Does a match mean “the best”?

It means a documented fit worth evaluating. We do not have a current, comparable independent quality benchmark across this whole catalogue. An agent's model, reasoning settings, tools, permissions and task change its results. GitHub activity, stars and a long feature list cannot establish a winner.

### How should I make the final choice?

Compare tools that fit your workflow, inspect their limitations, then run the same representative task with the same model and spending limit where possible. Check the resulting changes, permission prompts, recovery from errors and total usage. For the underlying model, use our [model rankings](/) and [plan comparison](/plans/).

04 / PROVENANCE

## More than one registry.

Research reviewed 2026-09-21. Product documentation can change after this snapshot.

 [Vercel Skills registry ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts)

Agent installation targets and skill paths. Registry inclusion does not establish product quality or current availability.

 Checked 2026-09-21
 [Agent Client Protocol registry ↗](https://cdn.agentclientprotocol.com/registry/v1/latest/registry.json)

Agents and distributions that connect to ACP clients. Wrappers and adapters are checked before counting products.

 Checked 2026-09-21
 [MCP extension client matrix ↗](https://github.com/modelcontextprotocol/modelcontextprotocol/blob/main/docs/extensions/client-matrix.mdx)

Official-hosted community matrix of opt-in client extensions. This is a discovery source, not an exhaustive quality leaderboard.

 Checked 2026-09-21

### Scope and limits

This is a researched index, not a claim to list every AI product. It focuses on tools you use to work with models: terminal agents first, alongside editors, graphical clients and related platforms. Documentation gaps stay visible. Retired, unmaintained and unresolved entries remain searchable, but cannot satisfy recommendations.

“Multiple providers” does not mean every model or endpoint works. “Local connection” does not promise offline operation, local execution of every feature, or private telemetry. An open-source CLI licence may not cover the same brand's hosted or desktop product. Check each profile's scope.

Research notes are in English. Device OS detection stays in your browser; a manual choice is saved locally and can be reset.

 Discovery entries that are not separate tools (2) - **Qoder CN:** Renamed to Qoder CN; counted once under its current product identity. [Source ↗](https://www.alibabacloud.com/help/en/lingma/introduction-of-lingma)
- **Universal:** A generic skills installation target, not a separately identified AI tool. [Source ↗](https://github.com/vercel-labs/skills/blob/main/src/agents.ts)

## Continue your investigation

 - [Find reusable skills](/skills/)
- [Inspect agents](/agents/)
- [Inspect tool connections](/mcp-servers/)
- [Check a repository](/check-repo/)