mcp-npx-fetch and defuddle-fetch-mcp-server
These two tools are competitors, as both aim to provide an MCP server for fetching and transforming web content, with "tokenizin-agency/mcp-npx-fetch" offering broader format transformations while "domdomegg/defuddle-fetch-mcp-server" specifically focuses on cleaning content for LLMs.
About mcp-npx-fetch
tokenizin-agency/mcp-npx-fetch
A powerful MCP server for fetching and transforming web content into various formats (HTML, JSON, Markdown, Plain Text) with ease.
Exposes four specialized fetch tools (fetch_html, fetch_json, fetch_txt, fetch_markdown) that handle content retrieval and format conversion, using JSDOM for HTML parsing and Turndown for HTML-to-Markdown transformation. Built on the Model Context Protocol SDK with TypeScript and Zod for runtime validation, supporting custom headers for authenticated requests. Integrates directly with Claude Desktop via configuration or runs as a standalone CLI tool installable via npm.
About defuddle-fetch-mcp-server
domdomegg/defuddle-fetch-mcp-server
🌐✨ Fetch content from the web, cleaned up nicely for LLMs (using Defuddle)
Implements the Model Context Protocol (MCP) as a drop-in replacement for the default fetch server, substituting Defuddle's site-specific extractors for Mozilla's Readability to achieve superior content extraction on modern webpages. Exposes a single `fetch` tool that returns cleaned markdown with rich metadata (title, author, publication date, word count), plus optional pagination via `start_index` and `max_length` parameters. Designed for seamless integration into Claude Desktop and other MCP-compatible AI assistants.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work