MCP.so
ログイン

airflow-mcp-server: An MCP Server for controlling Airflow

@MCP-Mirror

リポジトリ

airflow-mcp-server: An MCP Server for controlling Airflow について

Mirror of

#airflow#mcp-server#api-controller

基本情報

カテゴリ

データと分析

ライセンス

MIT license

ランタイム

python

トランスポート

stdio

公開者

MCP-Mirror

設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "airflow-mcp-server": {
      "command": "uvx",
      "args": [
        "airflow-mcp-server",
        "--base-url",
        "http://localhost:8080",
        "--auth-token",
        "<jwt_token>"
      ]
    }
  }
}

ツール

ツールは検出されませんでした

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is airflow-mcp-server?

An MCP (Model Context Protocol) server that controls Apache Airflow by interacting with Airflow’s REST APIs. It automatically parses the Airflow OpenAPI specification and supports both read-only and full‑access modes.

How to use airflow-mcp-server?

Install and run the server via uvx with the airflow-mcp-server CLI, providing the Airflow base URL and a JWT token. Optionally set --safe for read‑only operations or --unsafe (default) for all operations. Configure the server in your MCP client (e.g., Claude Desktop) using the mcpServers block.

Key features of airflow-mcp-server

  • Supports Airflow 3.0 and its JWT authentication.
  • Safe mode restricts operations to GET requests only.
  • Automatically fetches OpenAPI spec from the Airflow base URL.
  • Reads Airflow configuration to respect the page limit setting.
  • Exposes all Airflow API endpoints as MCP tools with proper descriptions.

Use cases of airflow-mcp-server

  • List and inspect DAGs, tasks, and their statuses without modifying anything.
  • Trigger DAG runs, clear task instances, or pause/unpause DAGs (in unsafe mode).
  • Monitor Airflow health and configuration from an AI‑assistant interface.
  • Safely explore an Airflow instance’s API surface through language model tools.

FAQ from airflow-mcp-server

What authentication does airflow-mcp-server require?

Only JWT token authentication is supported. Provide the token via the --auth-token parameter or the AUTH_TOKEN environment variable. Cookie and basic auth are not supported in Airflow 3.0.

What is the difference between safe and unsafe mode?

Safe mode allows only read‑only (GET) operations, preventing any modifications to your Airflow instance. Unsafe mode permits all operations, including creates, updates, and deletes.

How should I specify the base URL?

Set --base-url to the root Airflow URL (e.g., http://localhost:8080). Do not include /api/v2; the server automatically fetches the OpenAPI spec from ${base_url}/openapi.json.

Can I change the maximum number of items returned?

Yes. The default page limit is 100 items. You can change it by setting maximum_page_limit in the [api] section of your airflow.cfg file.

Is Airflow 3.0 supported?

Yes. The server is designed for Airflow 3.0 readiness and works exclusively with its JWT‑based authentication.

コメント

「データと分析」の他のコンテンツ

Bright Data MCP

brightdata-com

A powerful Model Context Protocol (MCP) server that provides an all-in-one solution for public web access.

2.5K

Salesforce MCP Server

tsmztech

Salesforce MCP Server

161

HubSpot MCP Server

baryhuang

A Model Context Protocol (MCP) server that enables AI assistants to interact with HubSpot CRM data, providing built-in vector storage and caching mechanisms help overcome HubSpot API limitations while improving response times.

128

🎓 Semantic Scholar MCP Server

JackKuo666

🔍 This project implements a Model Context Protocol (MCP) server for interacting with the Semantic Scholar API. It provides tools for searching papers, retrieving paper and author details, and fetching citations and references.

75

Google Analytics MCP Server

surendranb

Google Analytics 4 data to AI agents, agentic workflows, and MCP clients. Give agents analysis-ready access to website traffic, user behavior, and performance data with schema discovery, server-side aggregation, and safe defaults that reduce data wrangling.

225