Skip to main content
Glama

airflow-mcp-server

README.md2.15 kB
# airflow-mcp-plugin Airflow 3 plugin that mounts `airflow-mcp-server` as a Streamable HTTP endpoint at `/mcp` on the Airflow API server. Requirements: - Apache Airflow >= 3.0 (FastAPI backend) - Python >= 3.10 Install (recommended via main package extra): ```bash pip install "airflow-mcp-server[airflow-plugin]" ``` Or install the plugin directly: ```bash pip install airflow-mcp-plugin ``` Deploy: - Install into the Airflow webserver container environment (Docker/Compose/Helm) - Restart the webserver; Airflow auto-loads the plugin via entry point Use (stateless): - Endpoint: `http(s)://<airflow-host>/mcp` - Every request must include header: `Authorization: Bearer <access-token>` - The token is forwarded per-request to Airflow APIs (no shared auth state) - Mode per-request: - Safe (default): `http(s)://<airflow-host>/mcp` - Unsafe: `http(s)://<airflow-host>/mcp/?mode=unsafe` (enables POST/PUT/DELETE/PATCH) - Streamable HTTP (stateless) ### Usage with Claude Desktop Claude Desktop requires a helper wrapper to forward headers. Add the plugin endpoint via [`mcp-remote`](https://github.com/geelen/mcp-remote) inside `claude_desktop_config.json`: ```json { "mcpServers": { "airflow-mcp-plugin": { "command": "npx", "args": [ "mcp-remote", "http://localhost:8080/mcp", "--header", "Authorization:${MCP_AIRFLOW_TOKEN}" ], "env": { "MCP_AIRFLOW_TOKEN": "Bearer <access-token>" } } } } ``` See [`CONFIG.md`](CONFIG.md) for additional MCP client configuration examples. All clients must supply the `Authorization: Bearer <token>` header and connect over HTTP. ### Usage with VS Code Add the server definition to your `settings.json` under the `mcp.servers` block: ```json "mcp": { "servers": { "airflow-mcp-plugin": { "type": "http", "url": "http://localhost:8080/mcp/", "headers": { "Authorization": "Bearer <access-token>" } } } } ``` This configuration uses VS Code's native HTTP transport with per-request `Authorization` headers; update the URL if your Airflow instance is hosted elsewhere.

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/abhishekbhakat/airflow-mcp-server'

If you have feedback or need assistance with the MCP directory API, please join our Discord server