---
title: "Pinecone quickstart: review, setup & limitations · Undominated.ai"
canonical: https://undominated.ai/skills/pinecone-quickstart/
description: "Pinecone's onboarding skill: either a semantic-search demo on a new integrated index through the Pinecone MCP and a bundled upsert script, or a Pinecone Assistant built from the user's own documents."
---

# Pinecone quickstart: review, setup & limitations · Undominated.ai

> Pinecone's onboarding skill: either a semantic-search demo on a new integrated index through the Pinecone MCP and a bundled upsert script, or a Pinecone Assistant built from the user's own documents.

[← Explore all skills](/skills/)

DATABASES / Pinecone

# Pinecone quickstart

Pinecone's onboarding skill: either a semantic-search demo on a new integrated index through the Pinecone MCP and a bundled upsert script, or a Pinecone Assistant built from the user's own documents.

 [See setup guidance ↓](#setup)[Original source ↗](https://docs.pinecone.io/guides/get-started/quickstart)

SOURCE REVIEW

 Reviewed 2026-10-21
 Evidence 9 linked sources
 Publisher Pinecone
 Licence [MIT ↗](https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/LICENSE)
 Revision 885dc539fbc2
 [Read what was—and wasn’t—checked ↓](#review)

“Interactive Pinecone quickstart for new developers. Choose between two paths - Database (create an integrated index, upsert data, and query using Pinecone MCP + Python) or Assistant (create a Pinecone Assistant for document Q&A). Use when a user wants to get started with Pinecone for the first time or wants a guided tour of Pinecone's tools.”

 [Pinecone · upstream description ↗](https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/skills/pinecone-quickstart/SKILL.md) Our analysis follows below.

01 / THE REASONING

## Why this made the selection.

 - Path A is a fixed, checkable loop: list-indexes to test the key, create-index-for-model for a serverless index quickstart-skills (aws, us-east-1, llama-text-embed-v2, text mapped to chunk_text), a bundled upsert of nine records with pinecone==9.1.0, search-records queries that share no keywords with the records, and optional bge-reranker-v2-m3 reranking.
- Path B reuses the sibling pinecone-assistant scripts to create an assistant, upload .pdf, .md, .txt or .docx files and ask a question that is answered with citations, so both Pinecone products start from one skill.
- Pinecone publishes it in pinecone-io/skills, which the README calls Pinecone's official Agent Skills library, under the MIT License.

### A good fit for

 - Create a first integrated index and upsert nine sample records to see semantic search work
- Compare plain semantic search with bge-reranker-v2-m3 reranking on the same index
- Copy a standalone script that creates an index, upserts, searches and reranks in one run
- Set up a Pinecone Assistant over local documents and ask it a question with citations

### Weigh up before choosing

 - The skill directory skills/pinecone-quickstart last changed in a commit dated 2026-08-10 (git log at the pin); the pinned revision 885dc539 is dated 2026-09-18.
- Needs a Pinecone account and PINECONE_API_KEY; the README says the account is free at app.pinecone.io. Path A also needs the Pinecone MCP server configured, and the bundled scripts need uv.
- The SKILL.md at this path contains the literal snippet markers `<<clarify_style>>` and `<<api_key_setup>>` (twice). targets/README.md says tools/build.py fills these per publish target and that a skill published with `<<api_key_setup>>` visible is worse than either wording; the README's `npx skills add pinecone-io/skills` line names this repository, whose skills/ tree is the unbuilt base.
- Path B sends the user's local documents to Pinecone: upload.py uploads every file under --source that matches its default patterns (**/*.md, **/*.txt, **/*.pdf, **/*.docx, **/*.json), with metadata that includes each file's relative path, size and modification time. SKILL.md names only .pdf, .md, .txt and .docx.
- Path A creates a serverless index (aws, us-east-1) and Path B creates an assistant; the files read state no price for either. Removal is manual: troubleshooting offers `pc index delete -n quickstart-skills` through the Pinecone CLI, and SKILL.md gives no step for deleting the assistant or the second index, quickstart, that the standalone quickstart_complete.py creates.
- The README sends Claude Code users to a separate plugin repository, pinecone-io/pinecone-claude-code-plugin, which targets/claude-code.yaml names as the build target; that build adds allowed-tools "Skill(pinecone:assistant *), Bash, Read", which the base file does not declare.
- LICENSE is the MIT License with Copyright (c) 2026 Pinecone.

02 / THE REVIEW RECORD

## What we actually inspected.

Source review has boundaries. A clear record is more useful than a “safe” badge.

### Material inspected

 - skills/pinecone-quickstart/SKILL.md
- skills/pinecone-quickstart/scripts/upsert.py
- skills/pinecone-quickstart/scripts/quickstart_complete.py
- skills/pinecone-assistant/scripts/create.py
- skills/pinecone-assistant/scripts/upload.py
- README.md
- LICENSE
- targets/README.md
- targets/claude-code.yaml
- targets/codex.yaml
- targets/cursor.yaml
- The git tree at 885dc539, which holds nine skills/*/SKILL.md files

### Our findings

 - Path A uses Pinecone MCP tool calls plus a bundled Python upsert script; Path B delegates to the sibling pinecone-assistant skill's create, upload and chat scripts.
- Both bundled quickstart scripts pin pinecone==9.1.0 in uv inline metadata and stop if PINECONE_API_KEY is unset.
- upload.py's default patterns include **/*.json, which SKILL.md's list of accepted file types does not mention.
- The base SKILL.md holds unfilled snippet markers; targets/claude-code.yaml fills clarify_style with an instruction to use the AskUserQuestion tool and api_key_setup with a shell export line.
- The scripts pass source_tag values to the Pinecone client, and create.py stores agentic-ide-source metadata on each assistant it creates.

### Not established by this review

 - Not executed: it was not installed or run.
- No index or assistant was created, no records or files were uploaded, and no search or rerank was run.
- The Pinecone MCP server, the Claude Code plugin repository and the docs.pinecone.io pages were not read.

The review applies to the material and revision named here. A newer upstream release can change its behavior.

03 / PUT IT TO WORK

## Add a skill to your workflow.

[Upstream setup instructions ↗](https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/README.md)
 - The README prints `npx skills add pinecone-io/skills`. It installs all nine SKILL.md skills under skills/ at this revision, not this skill alone.
- For Claude Code the README says to try Pinecone's plugin at https://github.com/pinecone-io/pinecone-claude-code-plugin instead; that repository was not reviewed here.
- Before use, create an API key and run `export PINECONE_API_KEY="your-key"`, as the README prerequisites say. The README lists the Pinecone MCP server and uv as optional; this skill's Path A calls the Pinecone MCP tools, and both paths run their scripts with `uv run`.

### Before you start

 - A Pinecone account and API key
- The Pinecone MCP server for Path A
- uv to run the bundled scripts, which declare requires-python >=3.10

### Compatibility

Cursor · GitHub Copilot · Windsurf · Gemini CLI

### Implementation

Python

### Pinecone API key; creates an index or an assistant and uploads local files

 - Reads PINECONE_API_KEY from the environment, or from a .env file when the scripts run with `uv run --env-file .env`.
- Calls the Pinecone MCP tools list-indexes, create-index-for-model and search-records (with an optional rerank block) in the user's Pinecone project.
- Runs `uv run scripts/upsert.py --index quickstart-skills`, which upserts nine fixed records into namespace example-namespace with the pinecone==9.1.0 client.
- On request, copies scripts/quickstart_complete.py to ./pinecone_quickstart.py; that script creates a second index, quickstart, if it does not exist, then upserts, searches and reranks. No removal step is given for that index.
- Path B writes three sample markdown files to ./sample-docs/ when the user has none, runs ../pinecone-assistant/scripts/create.py to create an assistant, runs upload.py to upload the user's local files to it, and runs `uv run ../pinecone-assistant/scripts/chat.py --assistant my-assistant --message …` to send a question to the assistant.
- The scripts send source_tag values to Pinecone (pinecone_skills:upsert, pinecone_skills:index_quickstart, pinecone_skills:assistant), and create.py stores the metadata agentic-ide-source: pinecone-skills on the assistant.
- Troubleshooting offers deleting the index with `pc index delete -n quickstart-skills` through the Pinecone CLI.
- The base file declares no allowed-tools; the Claude Code build adds Skill(pinecone:assistant *), Bash and Read.

### Cost model

MIT for the skill files. The README says a Pinecone account is free at app.pinecone.io; the files read state no price for indexes, embedding, reranking, assistants or uploads.

04 / FOLLOW THE EVIDENCE

## The source trail.

Our notes are separate from the original resource. Check upstream before adopting a new version.

 - [pinecone-quickstart SKILL.md at 885dc539 ↗](https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/skills/pinecone-quickstart/SKILL.md) Checked 2026-10-21 https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/skills/pinecone-quickstart/SKILL.md Supports: upstreamDescription, homepage, whySelected, bestFor, limitations[1], limitations[2], limitations[3], limitations[4], access.permissions, install.prerequisites, review.findings
- [Repository README at 885dc539 ↗](https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/README.md) Checked 2026-10-21 https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/README.md Supports: compatibility, install, whySelected, limitations[1], limitations[2], limitations[5], access.cost, review.findings
- [MIT LICENSE at 885dc539 ↗](https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/LICENSE) Checked 2026-10-21 https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/LICENSE Supports: license, whySelected, limitations[6]
- [Bundled upsert script at 885dc539 ↗](https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/skills/pinecone-quickstart/scripts/upsert.py) Checked 2026-10-21 https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/skills/pinecone-quickstart/scripts/upsert.py Supports: whySelected, language, access.permissions, install.prerequisites, review.findings
- [Bundled standalone quickstart script at 885dc539 ↗](https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/skills/pinecone-quickstart/scripts/quickstart_complete.py) Checked 2026-10-21 https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/skills/pinecone-quickstart/scripts/quickstart_complete.py Supports: bestFor, limitations[4], access.permissions, review.findings
- [pinecone-assistant create.py at 885dc539 ↗](https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/skills/pinecone-assistant/scripts/create.py) Checked 2026-10-21 https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/skills/pinecone-assistant/scripts/create.py Supports: access.permissions, review.findings
- [pinecone-assistant upload.py at 885dc539 ↗](https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/skills/pinecone-assistant/scripts/upload.py) Checked 2026-10-21 https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/skills/pinecone-assistant/scripts/upload.py Supports: limitations[3], access.permissions, review.findings
- [Build target manifest guide (targets/README.md) at 885dc539 ↗](https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/targets/README.md) Checked 2026-10-21 https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/targets/README.md Supports: limitations[2], review.findings
- [Claude Code build manifest at 885dc539 ↗](https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/targets/claude-code.yaml) Checked 2026-10-21 https://github.com/pinecone-io/skills/blob/885dc539fbc2c777dee9764a5b05656663b43c5f/targets/claude-code.yaml Supports: limitations[5], access.permissions, review.findings

KEEP COMPARING

## Other approaches to consider.

Related by category or shared topics. These are alternatives to inspect, not a measured quality order.

 [### Turso Database ↗ Turso's skill for its in-process SQLite-compatible engine: hard constraints first, then decision trees to feature and SDK reference files, plus CLI, SQL and MCP quick references.](/skills/tursodatabase-turso-db/)[### Weaviate ↗ Weaviate skill that runs bundled Python scripts against a Weaviate Cloud cluster to list, inspect and create collections, search them, ask the Query Agent, and import files or example data.](/skills/weaviate-weaviate/)[### Azure Cosmos DB for Python ↗ Use Cosmos DB NoSQL clients for partition-aware document operations and parameterized queries.](/skills/microsoft-azure-cosmos-py/)

 [AI Tools ↗](/tools/)[Skills ↗](/skills/)[Agents ↗](/agents/)[MCP Servers ↗](/mcp-servers/)[Workflows ↗](/workflows/)

## Continue your investigation

 - [Choose a toolkit](/tools/)
- [Inspect execution hosts](/agents/)
- [Check tool access](/mcp-servers/)
- [Check project dependencies](/check-repo/)
