FastAPI

You run a FastAPI MCP Server to inspect and interact with FastAPI applications. It exposes practical tools to load apps, explore routes, view schemas, and analyze dependencies, making it easier to understand, document, and test your FastAPI projects from a centralized MCP client.

How to use

You connect to the MCP server from your MCP client to perform discovery, inspection, and debugging tasks. You can load your FastAPI app, list all routes, view route details, extract the OpenAPI schema, and inspect models. You can also search routes, analyze dependency injection usage, and view the source code of route handlers. These capabilities help you understand the API surface, generate documentation, and diagnose issues.

How to install

Prerequisites you need before installing and running the server:

• Python 3.10 or higher

• pip for installing Python packages

How to install

# From your project directory, install the MCP package in editable mode
pip install -e .

How to install

Run the MCP server locally in standalone mode to start serving MCP requests from your FastAPI project.

# Start the MCP server for your FastAPI project
python -m fastapi_mcp_server.server

Claude Desktop configuration (MCP client setup)

{
  "mcpServers": {
    "fastapi": {
      "command": "python",
      "args": [
        "-m",
        "fastapi_mcp_server.server"
      ],
      "cwd": "/path/to/your/fastapi/project"
    }
  }
}

Standalone usage

To run directly from your FastAPI project directory, execute the following command. This starts the server in the foreground, ready to accept MCP client connections.

python -m fastapi_mcp_server.server

Additional configuration and notes

• The MCP server exposes a set of tools for introspecting FastAPI applications, including loading apps, listing routes, retrieving route details, accessing the OpenAPI schema, listing models, and viewing individual model schemas.

Security and maintenance

• Run the MCP server in a controlled environment and limit access to trusted clients. If you expose it over a network, consider adding authentication or network access controls. Regularly update the MCP package to benefit from new introspection features and fixes.

Troubleshooting

• If the server fails to start, verify that you have Python 3.10+ installed and that the required package is installed in editable mode. Check that the working directory points to your FastAPI project when using Claude Desktop.

Available tools

load_fastapi_app

Load a FastAPI application from a module path so you can query routes, models, and OpenAPI.

list_routes

List all routes with optional filtering by HTTP method or tags to quickly understand the API surface.

get_route_details

Fetch comprehensive details about a specific route, including parameters, dependencies, and responses.

get_openapi_schema

Extract the OpenAPI schema for the loaded application or for selected paths.

list_models

List all Pydantic models used in the app to understand data structures.

get_model_schema

Get a detailed schema for a specific model, including fields and validation rules.

search_routes

Search routes by text query and criteria like authentication or tags.

analyze_dependencies

Analyze dependency injection usage to reveal how dependencies are resolved.

get_route_source

View the source code of route handlers for debugging and understanding implementations.