Fire in da houseTop Tip:Paying $100+ per month for Perplexity, MidJourney, Runway, ChatGPT and other tools is crazy - get all your AI tools in one site starting at $15 per month with Galaxy AI Fire in da houseCheck it out free

airflow-mcp-server

MCP.Pizza Chef: abhishekbhakat

The airflow-mcp-server is an MCP server that integrates Apache Airflow with the Model Context Protocol, allowing real-time control and interaction with Airflow workflows through its APIs. It facilitates secure, scoped access to Airflow's OpenAPI endpoints, enabling LLMs and AI agents to manage DAGs, trigger tasks, and monitor workflows programmatically within an MCP-enabled environment.

Use This MCP server To

Trigger Airflow DAG runs via natural language commands Monitor Airflow task statuses in real time Fetch Airflow DAG and task metadata for analysis Automate workflow management using LLM-driven agents Integrate Airflow control into AI-enhanced developer tools

README

airflow-mcp-server: An MCP Server for controlling Airflow

Find on Glama

Overview

A Model Context Protocol server for controlling Airflow via Airflow APIs.

Demo Video

airflow_mcp_server_demo.mp4

Setup

Usage with Claude Desktop

{
    "mcpServers": {
        "airflow-mcp-server": {
            "command": "uvx",
            "args": [
                "airflow-mcp-server",
                "--base-url",
                "http://localhost:8080",
                "--auth-token",
                "<jwt_token>"
            ]
        }
    }
}

Note:

  • Set base_url to the root Airflow URL (e.g., http://localhost:8080).
  • Do not include /api/v2 in the base URL. The server will automatically fetch the OpenAPI spec from ${base_url}/openapi.json.
  • Only JWT token is required for authentication. Cookie and basic auth are no longer supported in Airflow 3.0.

Operation Modes

The server supports two operation modes:

  • Safe Mode (--safe): Only allows read-only operations (GET requests). This is useful when you want to prevent any modifications to your Airflow instance.
  • Unsafe Mode (--unsafe): Allows all operations including modifications. This is the default mode.

To start in safe mode:

airflow-mcp-server --safe

To explicitly start in unsafe mode (though this is default):

airflow-mcp-server --unsafe

Considerations

Authentication

  • Only JWT authentication is supported in Airflow 3.0. You must provide a valid AUTH_TOKEN.

Page Limit

The default is 100 items, but you can change it using maximum_page_limit option in [api] section in the airflow.cfg file.

Tasks

  • Airflow 3 readiness
  • Parse OpenAPI Spec
  • Safe/Unsafe mode implementation
  • Parse proper description with list_tools.
  • Airflow config fetch (specifically for page limit)
  • Env variables optional (env variables might not be ideal for airflow plugins)

airflow-mcp-server FAQ

How do I configure the airflow-mcp-server base URL?
Set the base URL to your Airflow root URL without the /api/v2 suffix; the server auto-fetches the OpenAPI spec.
What authentication methods does airflow-mcp-server support?
It supports JWT token authentication to securely access Airflow APIs.
Can airflow-mcp-server be used with multiple LLM providers?
Yes, it is provider-agnostic and works with OpenAI, Claude, Gemini, and others.
How does airflow-mcp-server handle API versioning?
It automatically fetches and adapts to the Airflow OpenAPI spec for the configured version.
Is it possible to monitor Airflow task progress through this server?
Yes, the server exposes task status and metadata for real-time monitoring.
What platforms can run airflow-mcp-server?
It can run on any platform supporting Python and network access to Airflow's API endpoint.
How do I secure communication between the MCP client and airflow-mcp-server?
Use HTTPS and JWT tokens to ensure secure, authenticated communication.
Can I extend airflow-mcp-server to support custom Airflow plugins?
Yes, by modifying the server to expose additional API endpoints or custom logic.