Summary
One page, four shapes.
Reading a page and calling an API stop being different jobs: the format is chosen at the call — Markdown to read, text to summarize, JSON to work with — rather than cleaned up afterwards.
What it is
A fetcher that returns a URL in the shape you asked for: raw HTML, parsed JSON, plain text with tags and scripts stripped out, or the page converted to Markdown. Custom headers are optional on every call.
What you get
- A page returned as raw HTML
- A JSON endpoint or file returned already parsed
- A page stripped of markup, scripts and styles down to readable text
- A page converted to Markdown
- Custom headers on any of them, for endpoints that need one
Requirements
Node to build and run it, or an install through Smithery. It holds nothing and needs no credentials; whatever you point it at has to be reachable from the machine it runs on.
Setup effort
Build from source — clone the repository and build it, then point your client at the binary