/ 目录 / 演练场 / Airflow API MCP
● 社区 call518 ⚡ 即开即用

Airflow API MCP

作者 call518 · call518/MCP-Airflow-API

Control Apache Airflow with natural language via MCP — 43+ tools for Airflow 2.x and 3.0+.

Airflow API MCP is a MCP server that control apache airflow with natural language via mcp — 43+ tools for airflow 2.x and 3.0+ Use it from Claude Code, Cursor, Codex, or any MCP-compatible / skills-compatible agent.

为什么要用

核心特性

实时演示

实际使用效果

mcp-airflow-api.replay ▶ 就绪
0/0

安装

选择你的客户端

~/Library/Application Support/Claude/claude_desktop_config.json  · Windows: %APPDATA%\Claude\claude_desktop_config.json
{
  "mcpServers": {
    "mcp-airflow-api": {
      "command": "uvx",
      "args": [
        "mcp-airflow-api"
      ],
      "_inferred": true
    }
  }
}

打开 Claude Desktop → Settings → Developer → Edit Config。保存后重启应用。

~/.cursor/mcp.json · .cursor/mcp.json
{
  "mcpServers": {
    "mcp-airflow-api": {
      "command": "uvx",
      "args": [
        "mcp-airflow-api"
      ],
      "_inferred": true
    }
  }
}

Cursor 使用与 Claude Desktop 相同的 mcpServers 格式。项目级配置优先于全局。

VS Code → Cline → MCP Servers → Edit
{
  "mcpServers": {
    "mcp-airflow-api": {
      "command": "uvx",
      "args": [
        "mcp-airflow-api"
      ],
      "_inferred": true
    }
  }
}

点击 Cline 侧栏中的 MCP Servers 图标,然后选 "Edit Configuration"。

~/.codeium/windsurf/mcp_config.json
{
  "mcpServers": {
    "mcp-airflow-api": {
      "command": "uvx",
      "args": [
        "mcp-airflow-api"
      ],
      "_inferred": true
    }
  }
}

格式与 Claude Desktop 相同。重启 Windsurf 生效。

~/.continue/config.json
{
  "mcpServers": [
    {
      "name": "mcp-airflow-api",
      "command": "uvx",
      "args": [
        "mcp-airflow-api"
      ]
    }
  ]
}

Continue 使用服务器对象数组,而非映射。

~/.config/zed/settings.json
{
  "context_servers": {
    "mcp-airflow-api": {
      "command": {
        "path": "uvx",
        "args": [
          "mcp-airflow-api"
        ]
      }
    }
  }
}

加入 context_servers。Zed 保存后热重载。

claude mcp add mcp-airflow-api -- uvx mcp-airflow-api

一行命令搞定。用 claude mcp list 验证,claude mcp remove 卸载。

使用场景

实战用法: Airflow API MCP

Use Airflow API MCP for its core workflow

👤 Engineers and power users ⏱ ~15 min intermediate

何时使用: You want MCP server integration for data work without writing glue code.

前置条件
  • Repo installed / cloned — See repo README
步骤
  1. Install
    Set up the Airflow API MCP MCP server.✓ 已复制
    → MCP server listed in your agent config
  2. Invoke
    Use Airflow API MCP to run a representative task in data.✓ 已复制
    → Tool call succeeds; result returned

结果: Real data work completed via Airflow API MCP with one chat.

注意事项
  • Auth/credential setup is the most common blocker. — Follow the repo README for env vars and tokens.
搭配使用: filesystem

组合

与其他 MCP 搭配,撬动十倍杠杆

mcp-airflow-api + filesystem

Pair with filesystem for data workflows that need local file IO.

Use Airflow API MCP together with filesystem.✓ 已复制

工具

此 MCP 暴露的能力

工具输入参数何时调用成本
primary_call task-specific For the main use case Varies

成本与限制

运行它的成本

API 配额
Provider rate limits apply
每次调用 Token 数
Varies
费用
Free / pay provider directly
提示
Cache results; batch calls.

安全

权限、密钥、影响范围

凭据存储: Use environment variables or your OS keychain.
数据出站: Depends on integration

故障排查

常见错误与修复

Confirm install and restart your agent.

更多

资源

📖 阅读 GitHub 上的官方 README

🐙 查看未解决的 issue

🔍 浏览全部 400+ MCP 服务器和 Skills