---
title: "Chroma MCP: review, setup & limitations · Undominated.ai"
canonical: https://undominated.ai/mcp-servers/chroma/
description: "Lets an assistant create and query Chroma collections through local, HTTP or Chroma Cloud database connections."
---

# Chroma MCP: review, setup & limitations · Undominated.ai

> Lets an assistant create and query Chroma collections through local, HTTP or Chroma Cloud database connections.

[← Explore all mcp servers](/mcp-servers/)

VECTOR DATA MANAGEMENT / Chroma

# Chroma MCP

Lets an assistant create and query Chroma collections through local, HTTP or Chroma Cloud database connections.

 [See setup guidance ↓](#setup)[Original source ↗](https://github.com/chroma-core/chroma-mcp)

SOURCE REVIEW

 Reviewed 2026-10-14
 Evidence 4 linked sources
 Publisher Chroma
 Licence [Apache-2.0 ↗](https://github.com/chroma-core/chroma-mcp/blob/98ff67589bdcc31b730a5415ff9529433f949077/LICENSE)
 Revision 98ff67589bdc
 [Read what was—and wasn’t—checked ↓](#review)

“This is a MCP server for self-hosting your access to Chroma.”

 [Chroma · upstream description ↗](https://github.com/chroma-core/chroma-mcp/blob/98ff67589bdcc31b730a5415ff9529433f949077/README.md) Our analysis follows below.

01 / THE REASONING

## Why this made the selection.

 - Documents collection and document operations rather than claiming retrieval alone.
- Exposes persistent, ephemeral, HTTP and cloud Chroma connection choices in its configuration.

### A good fit for

 - Build a persistent small document collection for a project investigation.
- Query a selected collection with metadata filters and bounded result counts.

### Weigh up before choosing

 - The default client is ephemeral; persistence requires explicit configuration.
- Creation, modification and deletion tools are present; database authentication and embedding-provider choices affect data exposure.
- Embedding function configuration persists on collections created with Chroma 1.0+; older collections do not support that persistence behaviour.
- The default dotenv path is .chroma_env in the working directory; the working directory affects which credentials are loaded.

02 / THE REVIEW RECORD

## What we actually inspected.

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

### Material inspected

 - README.md (installation, capability and permission sections)
- pyproject.toml
- src/chroma_mcp/server.py
- LICENSE
- 2026-10-14 reread of retained pinned setup/configuration evidence for the specific operational details added in this batch.

### Our findings

 - The server entrypoint selects Chroma client and embedding configuration from CLI/environment settings.
- Chroma client types describe database connections, not four different MCP transports.
- Persistent mode needs --client-type persistent and an explicit --data-dir.
- Collection embedding choice is reused on retrieval; provider API keys are a separate dependency from Chroma database credentials.

### Not established by this review

 - Server startup, authentication and tool execution were not tested.
- No authenticated session, installation, permission bypass or package-to-source parity test was performed.

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

03 / PUT IT TO WORK

## Connect a server deliberately.

[Upstream setup instructions ↗](https://github.com/chroma-core/chroma-mcp/blob/98ff67589bdcc31b730a5415ff9529433f949077/README.md)

DOCUMENTED COMMAND

 uvx chroma-mcp Copy command ↗

Copying does not execute this command. It may retrieve a newer version than the reviewed source.

 - Select persistent, HTTP or cloud configuration if an ephemeral database is unsuitable.
- Configure database and optional embedding-provider credentials outside the prompt.
- The command above is the command and args of the README’s ephemeral Claude Desktop block (uvx with chroma-mcp), not a separate installer. It was not run.

### Before you start

 - uv/Python; database connection details and embedding-provider credentials when required.

### Documented configuration example

Use the client and authentication context described in the setup guidance. Configuration formats vary between hosts.

 "chroma": {
 "command": "uvx",
 "args": [
 "chroma-mcp"
 ]
} Copy configuration ↗

### Compatibility

An MCP client supporting stdio. · uv/Python; database connection details and embedding-provider credentials when required.

### Implementation

Python

### Transports

stdio

### Vector data management / Configured credentials or local access

 - Reads, adds, updates and deletes collection/document data.
- Selected remote embedding functions may send text to another service.

### Cost model

Local compute or Chroma Cloud usage and any external embedding service determine cost.

04 / FOLLOW THE EVIDENCE

## The source trail.

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

 - [Official documentation ↗](https://github.com/chroma-core/chroma-mcp/blob/98ff67589bdcc31b730a5415ff9529433f949077/README.md) Checked 2026-10-14 https://github.com/chroma-core/chroma-mcp/blob/98ff67589bdcc31b730a5415ff9529433f949077/README.md Supports: summary, whySelected, bestFor, limitations, install, access, compatibility, transports, configExample, review
- [Applicable source licence ↗](https://github.com/chroma-core/chroma-mcp/blob/98ff67589bdcc31b730a5415ff9529433f949077/LICENSE) Checked 2026-09-21 https://github.com/chroma-core/chroma-mcp/blob/98ff67589bdcc31b730a5415ff9529433f949077/LICENSE Supports: license
- [Implementation: pyproject.toml ↗](https://github.com/chroma-core/chroma-mcp/blob/98ff67589bdcc31b730a5415ff9529433f949077/pyproject.toml) Checked 2026-09-21 https://github.com/chroma-core/chroma-mcp/blob/98ff67589bdcc31b730a5415ff9529433f949077/pyproject.toml Supports: language, review
- [Implementation: src/chroma_mcp/server.py ↗](https://github.com/chroma-core/chroma-mcp/blob/98ff67589bdcc31b730a5415ff9529433f949077/src/chroma_mcp/server.py) Checked 2026-09-21 https://github.com/chroma-core/chroma-mcp/blob/98ff67589bdcc31b730a5415ff9529433f949077/src/chroma_mcp/server.py Supports: language, review

KEEP COMPARING

## Other approaches to consider.

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

 [### Pinecone MCP ↗ Search Pinecone documentation and manage or query indexes that use integrated embeddings.](/mcp-servers/pinecone/)[### Weaviate built-in MCP ↗ Inspect collections, search vector and keyword data, and optionally upsert objects through Weaviate’s built-in MCP endpoint.](/mcp-servers/weaviate/)[### Zilliz MCP Server ↗ Manages Zilliz Cloud projects and clusters and works with data in Milvus collections (create, insert, delete, vector and hybrid search) from an MCP client.](/mcp-servers/zilliz-cloud-mcp/)

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

## Continue your investigation

 - [Inspect resource evidence](/tools/)
- [Choose a host](/agents/)
- [Inspect reusable workflows](/skills/)
- [Use Undominated data](/api/)
