Airflow
Connects AI assistants to Apache Airflow instances via the Model Context Protocol (MCP) to enable inspection and management.
About
The Airflow MCP Server acts as a human-entrypoint, exposing a safe and focused set of tools that allow AI assistants to interact with Apache Airflow. It enables inspection of DAGs, runs, and logs, with optional write operations gated by client approval for enhanced security. Responses are structured JSON objects, simplifying integration with AI models and providing traceability through a `request_id`. The server supports configuring multiple Airflow instances and offers robust server-side filtering for logs, facilitating efficient incident triage and automated Airflow management workflows.
Key Features
- Includes gated write operations requiring client approval for sensitive actions.
- Exposes structured JSON API for inspecting and managing Airflow DAGs, runs, and logs.
- Validates Airflow UI URLs with an SSRF guard for enhanced security.
- Supports configurable connections to multiple Apache Airflow instances.
- 1 GitHub stars
- Provides server-side log filtering and truncation for efficient incident diagnosis.
Use Cases
- Automating incident response workflows for Airflow pipeline failures via AI assistants.
- Programmatic inspection of Airflow task metadata, rendered fields, and logs.
- Integrating AI-powered agents to monitor, query, and manage Airflow DAGs and runs.
Sponsored