chromadb fastapi mcp

Installation

1. Clone the repository
2. Install dependencies:

   uv sync
   

3. Copy .env.example to .env and configure settings:

   cp .env.example .env
   

4. Edit .env with your preferred settings

Configuration

Configure the following environment variables in your .env file:

• CHROMA_CLIENT_TYPE: Either ephemeral or persistent
• CHROMA_DATA_DIR: Directory for storage when using persistent client
• OPENAI_API_KEY: Your OpenAI API key for embeddings

Usage

Run the server:

cd /path/to/chromadb-fastapi-mcp
uv run python -m app.main

Or with uvicorn directly:

uv run uvicorn app.main:app --reload

The server will start at http://localhost:8000, and API documentation is available at http://localhost:8000/docs.

API Endpoints

Collections

• GET /api/collections/ - List all collections
• POST /api/collections/ - Create a new collection
• GET /api/collections/{collection_name}/peek - Peek at documents in a collection
• GET /api/collections/{collection_name}/info - Get collection information
• GET /api/collections/{collection_name}/count - Get document count in a collection
• PUT /api/collections/{collection_name} - Modify a collection
• DELETE /api/collections/{collection_name} - Delete a collection

Documents

• POST /api/documents/add - Add documents to a collection
• POST /api/documents/query - Query documents from a collection
• POST /api/documents/get - Get documents from a collection
• PUT /api/documents/update - Update documents in a collection
• DELETE /api/documents/delete - Delete documents from a collection

MCP server

https://github.com/tadata-org/fastapi_mcp

Connecting to the MCP Server using SSE

Once your FastAPI app with MCP integration is running, you can connect to it with any MCP client supporting SSE, such as Cursor:

1. Run your application.

2. In Cursor -> Settings -> MCP, use the URL of your MCP server endpoint (e.g., http://localhost:8000/mcp) as sse.

3. Cursor will discover all available tools and resources automatically.

Connecting to the MCP Server using mcp-proxy stdio

If your MCP client does not support SSE, for example Claude Desktop:

1. Run your application.

2. Install mcp-proxy, for example: uv tool install mcp-proxy.

3. Add in Claude Desktop MCP config file (claude_desktop_config.json):

On Windows:

{
  "mcpServers": {
    "my-api-mcp-proxy": {
        "command": "mcp-proxy",
        "args": ["http://127.0.0.1:8000/mcp"]
    }
  }
}

On MacOS:

Find the path to mcp-proxy by running in Terminal: which mcp-proxy.

{
  "mcpServers": {
    "my-api-mcp-proxy": {
        "command": "/Full/Path/To/Your/Executable/mcp-proxy",
        "args": ["http://127.0.0.1:8000/mcp"]
    }
  }
}

Find the path to mcp-proxy by running in Terminal: which uvx.

{
  "mcpServers": {
    "my-api-mcp-proxy": {
        "command": "/Full/Path/To/Your/uvx",
        "args": ["mcp-proxy", "http://127.0.0.1:8000/mcp"]
    }
  }
}

4. Claude Desktop will discover all available tools and resources automatically