Skip to main content
gbrain
Sign in

Page fetching · MCP

Page fetching in any assistant.

Turn a URL into readable text, including pages that need a browser.

For the agenda PDF, the council minutes, the pricing page that renders in JavaScript. One command in, text out, on the same record as everything else. 

claude mcp add gbrain https://gbrain.io/mcp --header "Authorization: Bearer <token>"

One endpoint for everything on the plane, not one per service. What this client can call is decided by its permission, not by the address.

The tools it gets

Read from the tool plane, so this is what an assistant is handed today, minus anything its permission does not allow.

firecrawl_search
Search the web (Firecrawl)
firecrawl_scrape
Read a page

Or from a terminal

The same Page fetching, four commands from a shell prompt, on the same permission and the same record.

Page fetching from the command line

How the whole thing fits

One permission, one daily ceiling, one record, across every client that asks.

Tools

Also on the plane