us-all/airflow-mcp-server

Overview

This tool is an MCP server that connects to an Airflow instance via its REST API. It exposes DAGs, runs, task instances, log tails, and operations like trigger and clear as MCP tools. This allows MCP clients to manage Airflow workflows directly through the server's interface.

Best for

Best for
Developers using MCP clients who need to trigger and inspect Airflow DAGs programmatically.

Use cases

• Trigger and monitor DAG runs from an MCP client
• Read task instances and tail logs without Airflow UI access
• Clear or trigger task instances as part of automated workflows

Notes

This tool is an MCP server that connects to an Airflow instance via its REST API. It exposes DAGs, runs, task instances, log tails, and operations like trigger and clear as MCP tools. This allows MCP clients to manage Airflow workflows directly through the server’s interface.

0 stars on GitHub. Last updated 2026-06-18. Licensed MIT.

Use cases

• Trigger and monitor DAG runs from an MCP client
• Read task instances and tail logs without Airflow UI access
• Clear or trigger task instances as part of automated workflows

Pros

• Direct integration with Airflow’s REST API for reliable operations
• Enables workflow control from any MCP-compatible client
• Uses TypeScript for type safety and maintainability

Cons

• No stars or visible community adoption yet
• Requires Airflow REST API to be exposed and accessible
• Only exposes operations available through the REST API

Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.