Fetch MCP
Local · stdioA reference web-content server: downloads a page and returns it as markdown, chunked to fit the model context. A minimal alternative to a full browser when only page text is needed.
Why this server
- Reads long documents page by page via start_index — never cuts off mid-document
- Respects robots.txt for model-initiated requests, but not for explicit user requests
- raw=true returns the original HTML when markdown conversion breaks the layout
- Minimal footprint: a single fetch tool instead of spinning up a full browser
Usage examples
Prompt: Read this blog post and summarize the key points
Result: The agent called fetch(url), got the markdown version of the page and pulled out the key points, no browser needed.
Prompt: The docs page is long — show me just the authentication section
Result: The agent paged through the document via start_index, locating the section without loading the whole text at once.
Local server
This MCP runs locally on the user's machine (stdio), so availability isn't tracked — rely on reviews and the rating instead.
Reviews
No reviews yet.