GeekNews MCP Server

A Model Context Protocol (MCP) server for retrieving and processing news data from GeekNews, designed for integration in the MCP server ecosystem. Implemented in Python and uses BeautifulSoup for web scraping.

Features

• Article Retrieval: Fetch articles from GeekNews with the get_articles tool.
  • Supports specifying article type (top, new, ask, show) and the number of articles to retrieve.
  • Each article includes title, URL, points, author, time, comment count, and rank.
  • Uses cached data to reduce server load.
• Weekly News Retrieval: Fetch weekly news from GeekNews with the get_weekly_news tool.
  • Specify a particular weekly news ID or get the latest weekly news.
  • Provides details such as title, number, ID, content, URL, and item list.
  • Each item includes title, URL, and rank.
  • Uses cached data for efficiency.
• Caching Mechanism:
  • Automatically parses and caches data once per day.
  • Valid cached data is used instead of making redundant requests to GeekNews.
  • Cache is valid for 24 hours and auto-renews as needed.
  • A scheduler checks cache validity and refreshes it as required.
• Implementation Details:
  • Python codebase with modular structure: models, parser, client, config, cache, scheduler, server, and entry point (main.py).
  • Cache data stored in the user's home directory under .cache/geeknews-mcp.
  • Can be installed and run locally or via Smithery MCP server ecosystem.

Installation & Usage

• Via Smithery: Add the server to your MCP configuration and run using Smithery CLI with your issued key.
• Local Installation: Clone the repo, set up the environment with uv, activate the virtual environment, and run the server for local testing.

Notes

• The server depends on the HTML structure of GeekNews. If the website changes, parsing logic may require updates.

License

MIT License

Source

https://github.com/the0807/GeekNews-MCP-Server

Pricing

• Open source, free to use under the MIT License.