Chroma MCP logo

Chroma MCP

by chroma-core

Connects AI clients to local, persistent, self-hosted, or cloud Chroma databases for vector retrieval and collection management.

Install Chroma MCP

Pick your client, then add it in one click or paste the config.

STDIOPython
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "chroma-core-chroma-mcp": {
      "command": "uvx",
      "args": [
        "chroma-mcp"
      ],
      "env": {
        "CHROMA_CLIENT_TYPE": "YOUR_VALUE_HERE",
        "CHROMA_DATA_DIR": "YOUR_VALUE_HERE",
        "CHROMA_TENANT": "YOUR_VALUE_HERE",
        "CHROMA_DATABASE": "YOUR_VALUE_HERE",
        "CHROMA_API_KEY": "YOUR_VALUE_HERE",
        "CHROMA_HOST": "YOUR_VALUE_HERE",
        "CHROMA_PORT": "YOUR_VALUE_HERE",
        "CHROMA_CUSTOM_AUTH_CREDENTIALS": "YOUR_VALUE_HERE"
      }
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Values are written into the snippet above so you can paste it as is. They never leave your browser.

We couldn’t confirm it starts correctly

The install command started, but didn't respond the way we expected when we tried to talk to it.

uvx chroma-mcp

initialize succeeded but no response to tools/list.

This is an experimental automated check and can have false negatives — missing environment variables, a slow cold install, etc. It doesn’t necessarily mean something’s wrong. Last checked 24d ago.

Overview

chroma-core/chroma-mcp MCP server exposes Chroma collections and documents through MCP tools for local, persistent, self-hosted HTTP, and Chroma Cloud clients. It supports collection administration, document insertion and updates, semantic queries, metadata filters, and document-content filters. Configure the client type with command-line arguments or environment variables, then use it when an agent needs to read or modify Chroma-backed data.

Use cases

  • Create collections for agent memory
  • Query indexed documents with semantic search
  • Filter retrieval results by metadata or content
  • Maintain documents and collection metadata
  • Connect agents to Chroma Cloud or self-hosted Chroma

Key features

  • Local, persistent, HTTP, and cloud Chroma clients
  • Collection lifecycle and metadata management
  • Semantic document queries
  • Metadata and document-content filtering
  • Document CRUD operations
  • Configurable embedding functions

Tools13

What Chroma MCP lets your agent do. Expand a tool for its parameters.

Verified live~2.8k tokens of context

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked Today

Recent health check succeeded.

21/21 checks healthy over the last 79d

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 9d ago via OSV.dev · chroma-mcp (PyPI)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars600Stargazers on the source repository.
Last commit1y agoMost recent push to the default branch.
Install checkInconclusiveStarted but did not complete the MCP handshake in our test.
Tools exposed13Callable tools this server registers over MCP.
Directory activity4 viewsConfig copies, upvotes, and views on AllMCPs.

How Chroma MCP works

What chroma-core/chroma-mcp does

The chroma-core/chroma-mcp MCP server connects an MCP-compatible AI client to Chroma, an embedding database. It supports ephemeral in-memory storage, persistent file-based storage, self-hosted Chroma over HTTP, and Chroma Cloud. This makes it suitable for agents that need retrieval over indexed documents, shared knowledge bases, or stored application data.

As a 🗄️ Databases entry, chroma-core/chroma-mcp focuses on database access rather than application-specific workflows. Agents can inspect collections, create or modify them, manage metadata, and add, retrieve, update, or delete documents.

How it works

The server runs as a standard MCP process and exposes Chroma operations as named tools. Collection tools cover listing with pagination, creation, peeking at sample documents, retrieving collection information and counts, renaming or changing metadata, forking, and deletion.

Document tools accept required IDs for inserts and updates where applicable. Queries can use semantic search with a requested result count, metadata predicates, and document-content predicates. Supported filter patterns include equality and comparisons, logical AND and OR, text containment, and regular expressions. Retrieval can also be limited with IDs, offsets, and limits.

Collections can be created with one of the supported embedding function names: default, cohere, openai, jina, voyageai, ollama, or roboflow. The selected collection configuration is reused for later inserts and retrievals when supported by the Chroma version.

Setup and configuration

Use the default invocation for an ephemeral client, or pass --client-type persistent and --data-dir for file-backed storage. A cloud connection uses --client-type cloud, tenant, database, and API key arguments. A self-hosted deployment uses --client-type http with host, port, optional custom authentication credentials, and SSL settings.

The same settings can come from system environment variables or a dotenv file. The default dotenv filename is .chroma_env; a different path can be selected with --dotenv-path or CHROMA_DOTENV_PATH. Command-line options take precedence over environment variables. External embedding providers require API keys using the CHROMA_<PROVIDER>_API_KEY naming pattern.

Tools and capabilities

chroma-core/chroma-mcp provides these MCP tools:

  • List, create, inspect, count, modify, fork, and delete collections.
  • Add documents with IDs and optional metadata.
  • Query documents with semantic search, metadata filters, and content filters.
  • Retrieve documents by IDs or filters with pagination.
  • Update document text, metadata, or embeddings.
  • Delete documents by ID.

Use chroma_peek_collection for a limited sample, while chroma_get_documents is intended for filtered or ID-based retrieval. Query and retrieval responses can request specific included fields; documented defaults include documents and metadata, with distances also included by default for queries.

Limitations and notes

External embedding functions depend on their corresponding API keys. Persisted embedding-function configuration requires Chroma version 1.0.0 or later; collections created with versions through 0.6.3 do not support that behavior. Update operations require at least one replacement field, and supplied update lists must match the number of IDs. Empty ID lists are invalid for update and delete operations.

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Chroma MCP

chroma-core/chroma-mcp is an MCP server that connects AI clients to local, persistent, self-hosted HTTP, or Chroma Cloud databases. Its main tools manage collections and documents, including semantic queries, filtered retrieval, inserts, updates, deletions, collection inspection, and collection metadata changes.

Run it through `uvx chroma-mcp` and add that command to an MCP client configuration.

Set the client type to `cloud` and provide the tenant, database, and API key. The server then connects to Chroma Cloud through `api.trychroma.com`; these values can be supplied as command-line arguments or environment variables.

The documented client configuration is for Claude Desktop, while the server itself uses a standard MCP process launched with `uvx`.

Related MCP servers

Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewChroma MCP AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/chroma-core-chroma-mcp?style=directory)](https://allmcps.com/mcp/chroma-core-chroma-mcp)
HTML Embed
<a href="https://allmcps.com/mcp/chroma-core-chroma-mcp"><img src="https://allmcps.com/api/badge/chroma-core-chroma-mcp?style=directory" alt="Chroma MCP on AllMCPs" /></a>