Airflow MCP Server
Airflow MCP server — read DAGs, runs, task instances, log tails; trigger and clear (write-gated).
What is Airflow MCP Server?
Airflow MCP Server is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to airflow mcp server — read dags, runs, task instances, log tails; trigger and clear (write-gated).
Airflow MCP server — read DAGs, runs, task instances, log tails; trigger and clear (write-gated).
This server falls under the Cloud Services category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- Airflow MCP server — read DAGs, runs, task instances, log ta
Use Cases
Maintainer
Works with
Installation
NPM
npx -y @us-all/airflow-mcpPIP
pip install airflow-mcp-serverManual Installation
npx -y @us-all/airflow-mcpConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use Airflow MCP Server
The Airflow MCP Server exposes Apache Airflow's REST API to AI assistants, enabling natural language control over DAGs, task instances, connections, and execution logs. It supports hierarchical tool discovery (browsing by category) as well as a static mode where all tools are immediately available, and includes a safe mode that restricts operations to read-only to prevent accidental workflow modifications. Platform engineers and data teams use it to monitor pipeline health, trigger DAG runs, and debug task failures without leaving their AI assistant.
Prerequisites
- Apache Airflow 3.0+ running with the REST API enabled (JWT authentication required)
- A valid JWT auth token for your Airflow instance
- Python 3.10+ and uvx (part of the uv toolchain) installed
- An MCP-compatible client such as Claude Desktop or Claude Code
Install uv to get uvx
The server is best run via uvx, which is bundled with the uv Python package manager.
curl -LsSf https://astral.sh/uv/install.sh | shObtain a JWT token from your Airflow instance
Generate or retrieve a JWT token from your Airflow 3.0+ instance. Airflow 3.0 uses JWT-only authentication for the REST API.
Test the server from the command line
Run the server directly to verify connectivity. The base URL should NOT include the /api/v2 suffix — the server appends it automatically.
uvx airflow-mcp-server --base-url http://localhost:8080 --auth-token <your-jwt-token>Add the server to your MCP client configuration
Open claude_desktop_config.json and add the Airflow MCP server block. Use --safe to restrict to read-only operations.
{
"mcpServers": {
"airflow-mcp-server": {
"command": "uvx",
"args": [
"airflow-mcp-server",
"--base-url",
"http://localhost:8080",
"--auth-token",
"<your-jwt-token>"
]
}
}
}Restart your client and explore DAG categories
Restart Claude Desktop. In hierarchical mode (the default), ask the assistant to list available tool categories, then drill into DAG management or task operations.
Airflow MCP Server Examples
Client configuration
Claude Desktop configuration for Airflow MCP server in safe (read-only) mode.
{
"mcpServers": {
"airflow-mcp-server": {
"command": "uvx",
"args": [
"airflow-mcp-server",
"--safe",
"--base-url",
"http://localhost:8080",
"--auth-token",
"<your-jwt-token>"
]
}
}
}Prompts to try
Example prompts for interacting with Apache Airflow through the MCP server.
- "List all DAGs in my Airflow instance and show which ones are currently paused"
- "Trigger a run of the etl_daily_pipeline DAG with today's date as the logical date"
- "Show me the last 5 runs of the data_ingestion DAG and their statuses"
- "Fetch the logs for the failed transform_data task in the most recent DAG run"
- "Clear and restart all failed tasks in the last run of the reporting_pipeline DAG"Troubleshooting Airflow MCP Server
Server fails to connect with a 401 or 403 error
Verify that your JWT token is still valid and that it was generated for the correct Airflow instance. Airflow 3.0 uses short-lived JWT tokens — regenerate if expired.
The base URL path causes 404 errors on API calls
The --base-url argument must NOT include /api/v2. Provide only the root URL, e.g., http://localhost:8080, and the server will append the correct API path automatically.
Tool modifications are blocked even though safe mode is not set
Check whether your Airflow instance has write operations disabled at the API level. If you need write access, ensure the JWT token has the required role permissions in Airflow's RBAC configuration.
Frequently Asked Questions about Airflow MCP Server
What is Airflow MCP Server?
Airflow MCP Server is a Model Context Protocol (MCP) server that airflow mcp server — read dags, runs, task instances, log tails; trigger and clear (write-gated). It connects AI assistants to external tools and data sources through a standardized interface.
How do I install Airflow MCP Server?
Install via npm with the command: npx -y @us-all/airflow-mcp. Then add the server configuration to your AI client's JSON config file (e.g., claude_desktop_config.json or .cursor/mcp.json).
Which AI clients work with Airflow MCP Server?
Airflow MCP Server works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is Airflow MCP Server free to use?
Yes, Airflow MCP Server is open source and available under the MIT License license. You can use it freely in both personal and commercial projects.
Airflow MCP Server Alternatives — Similar Cloud Services Servers
Looking for alternatives to Airflow MCP Server? Here are other popular cloud services servers you can use with Claude, Cursor, and VS Code.
Open WebUI
★ 138.2kUser-friendly AI Interface (Supports Ollama, OpenAI API, ...)
Anything LLM
★ 60.4kThe all-in-one AI productivity accelerator. On device and privacy first with no annoying setup or configuration.
LocalAI
★ 46.4kLocalAI is the open-source AI engine. Run any model - LLMs, vision, voice, image, video - on any hardware. No GPU required.
Nacos
★ 33.0kan easy-to-use dynamic service discovery, configuration and service management platform for building AI cloud native applications.
Xiaozhi ESP32
★ 26.7k本项目为xiaozhi-esp32提供后端服务，帮助您快速搭建ESP32设备控制服务器。Backend service for xiaozhi-esp32, helps you quickly build an ESP32 device control server.
Gateway
★ 11.8kA blazing fast AI Gateway with integrated guardrails. Route to 1,600+ LLMs, 50+ AI Guardrails with 1 fast & friendly API.
Browse More Cloud Services MCP Servers
Explore all cloud services servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up Airflow MCP Server in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use Airflow MCP Server?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.