antonio-mello-ai/mcp-airflow

What it does

mcp-airflow is an MCP (Model Context Protocol) server that provides programmatic access to Apache Airflow's capabilities through its REST API. It enables users to manage their data pipelines directly from AI assistants and other MCP-compatible clients, eliminating the need to switch between tools.

Key features

• List and inspect DAGs (Directed Acyclic Graphs)
• View DAG runs and task instance details
• Trigger new DAG runs programmatically
• Monitor failed DAGs and identify issues
• Check scheduler and metadatabase health status
• 7 specialized tools built with FastMCP framework

How to set up

Installation is straightforward using uvx package manager. Run uvx mcpairflow to install and initialize the MCP server. Configure it with your Apache Airflow REST API credentials and endpoint URL. Once set up, the server integrates seamlessly with Claude, ChatGPT, and other MCP-compatible platforms.

Best for

Data engineers and DevOps professionals who want to manage Airflow workflows through AI assistants, enabling natural language interaction with their data pipelines and reducing operational overhead.