MCP-Airflow-API

概述

MCP-Airflow-API 是一个全面的 MCP 服务器，将 Apache Airflow REST API 操作转换为自然语言工具，使用户能够通过直观命令而非复杂的 API 调用来管理 Airflow 集群。该项目支持 Airflow 2.x（43 个工具）和 3.0+（45+ 个工具），可通过环境变量动态选择版本，使其能够适应不同的 Airflow 部署而无需更改代码。

试试问 AI

装完之后，这里有 5 个你可以让 AI 做的事：

• list_dags

  Retrieve a list of all DAGs in Airflow

  试试:Show me all DAGs试试:List all available DAGs试试:What DAGs do I have?
• get_dag_details

  Get detailed information about a specific DAG

  试试:Show me details for the ETL DAG试试:What's the configuration of my daily pipeline?
• trigger_dag

  Trigger a DAG run for a specific DAG

  试试:Run the data processing DAG now试试:Trigger my ETL pipeline
• pause_dag

  Pause a DAG to prevent it from running

  试试:Pause the maintenance DAG试试:Stop the nightly reporting from running
• resume_dag

  Resume a paused DAG

  试试:Unpause the backup DAG试试:Enable the nightly reporting again
• get_dag_runs

  Get all runs for a specific DAG

  试试:Show me recent runs of the analytics DAG试试:When did the ETL pipeline last run?
• get_task_instances

  Get task instances for a specific DAG run

  试试:Show me the tasks from yesterday's run试试:What tasks were in the latest pipeline execution?
• get_task_log

  Get the log output for a specific task instance

  试试:Show me the log for the data quality check task试试:What went wrong with the extraction task?
• get_variables

  Retrieve variables stored in Airflow

  试试:Show me all variables试试:What's the value of the database URL variable?
• set_variable

  Create or update a variable in Airflow

  试试:Set the timeout variable to 3600试试:Create a new API key variable
• get_connections

  Get all connections configured in Airflow

  试试:Show me all database connections试试:What connections are configured?
• add_connection

  Add a new connection to Airflow

  试试:Add a new PostgreSQL connection试试:Create a connection to the warehouse database

可对比工具

安装

uvx --python 3.12 mcp-airflow-api

{
  "mcpServers": {
    "mcp-airflow-api": {
      "command": "uvx",
      "args": ["--python", "3.12", "mcp-airflow-api"],
      "env": {
        "AIRFLOW_API_VERSION": "v2",
        "AIRFLOW_API_BASE_URL": "http://localhost:8080/api",
        "AIRFLOW_API_USERNAME": "airflow",
        "AIRFLOW_API_PASSWORD": "airflow"
      }
    }
  }
}

git clone https://github.com/call518/MCP-Airflow-API.git
cd MCP-Airflow-API
cp .env.example .env
# 编辑 .env 文件，添加您的 Airflow API 设置
docker-compose up -d

FAQ

支持哪些 Airflow 版本？

MCP-Airflow-API 支持 Airflow 2.x（API v1，43 个工具）和 Airflow 3.0+（API v2，45+ 个工具）。版本通过 AIRFLOW_API_VERSION 环境变量选择。

如何处理身份验证？

它支持基本身份验证（用户名/密码）和流式 HTTP 模式的令牌身份验证。通过环境变量进行配置。

Hacker News 讨论

开发者社区最近的相关讨论。

• Show HN: MCP-Server for Control Airflow Cluster▲ 2
  帖子 by call518 · 2025-08-17

MCP-Airflow-API 对比