Submit Log in

Chroma

by Chroma

v0.2.6

Repository Docs

Databases & Data

vector-databaseembeddingsragchroma

Summary

Chroma's official MCP server — build and query vector collections directly from an AI client.

Features

Create collections and add documents from the assistant
Semantic, keyword, and metadata-filtered queries
Ephemeral, persistent, and cloud client modes

Installation

Set up this MCP server in your favorite AI agent — copy a ready-made configuration below.

Any MCP-compatible agent

Most agents (Claude, Cursor, Windsurf, VS Code, and more) read a standard mcpServers configuration.

Open your agent's MCP configuration file.
Merge the snippet below into it.
Restart the agent — the "Chroma" tools will be available.

{
  "mcpServers": {
    "chroma-mcp": {
      "command": "uvx",
      "args": [
        "chroma-mcp",
        "--client-type",
        "persistent",
        "--data-dir",
        "./chroma"
      ]
    }
  }
}

Claude Code

Add the server with a single CLI command.

claude mcp add chroma-mcp -- uvx chroma-mcp --client-type persistent --data-dir ./chroma

Claude Desktop

Add the server to claude_desktop_config.json (Settings → Developer → Edit Config).

{
  "mcpServers": {
    "chroma-mcp": {
      "command": "uvx",
      "args": [
        "chroma-mcp",
        "--client-type",
        "persistent",
        "--data-dir",
        "./chroma"
      ]
    }
  }
}

Cursor

{
  "command": "uvx",
  "args": [
    "chroma-mcp",
    "--client-type",
    "persistent",
    "--data-dir",
    "./chroma"
  ]
}

Windsurf

{
  "command": "uvx",
  "args": [
    "chroma-mcp",
    "--client-type",
    "persistent",
    "--data-dir",
    "./chroma"
  ]
}

VS Code

{
  "command": "uvx",
  "args": [
    "chroma-mcp",
    "--client-type",
    "persistent",
    "--data-dir",
    "./chroma"
  ]
}

Description

Chroma's official MCP server exposes the whole embedding-database workflow to an assistant: create a collection, add documents, and query it, all without writing code.

What it exposes

Collection management — create, list, inspect, and delete collections
Document operations — add, update, and delete, with automatic embedding
Querying by semantic similarity, full-text search, or metadata filter, and combinations of all three
Client modes — ephemeral for throwaway work, persistent for a local directory, HTTP for a running server, and cloud for Chroma Cloud

Why this is useful in practice

It closes the loop on retrieval experiments. Instead of writing an ingestion script, running it, and then testing retrieval in another script, an assistant can add documents and immediately query them, iterating on chunking and metadata in one conversation.

It also works as a durable memory store: give the assistant a persistent collection and it can accumulate knowledge across sessions.

Setup

uvx chroma-mcp --client-type persistent --data-dir ./chroma, or with cloud credentials for a hosted collection.

Related MCP Servers

MCP: Weaviate MCP Server

by Weaviate

New

Weaviate's MCP server, built into the database itself — hybrid search, schema inspection, tenant listing and batch upserts at /v1/mcp, scoped by RBAC.

Databases & Data

MCP: MCP Toolbox for Databases

by Google

New

Google's open-source MCP server for enterprise databases — one binary that fronts AlloyDB, BigQuery, Spanner, Postgres, MySQL, Oracle, MongoDB, Snowflake and more.

Databases & DataCloud & Infrastructure