Skip to content
Han AI

Page fetch

Han AI can fetch a URL directly, strip the HTML to clean readable text, and read it as context for whatever you asked next.

What it does

Pulls a raw HTTP response, removes script, style, and navigation chrome, and returns the visible text up to a 20,000 character cap.

FieldValue
Schema namefetch_page
Powered byNode fetch plus strip-tags
Output cap20,000 characters
API key requiredNo

When Han AI uses it

  • The user gives a URL and asks a question about it.
  • A web search result looks promising and needs to be read in full.
  • A document lives on a static page and does not require JavaScript to render.

Examples

  • “Read this and tell me what their cancellation policy actually says: https://…”
  • “Compare the warranty terms on these three product pages.”
  • “Pull the board members listed on their about page.”

Limits

  • Static HTML only. Pages that render content via JavaScript return mostly empty markup — Han AI falls back to the browser tool.
  • Pages over 20,000 characters are truncated. Han AI summarises what it can see and flags the truncation.
  • Login walls and bot challenges (Cloudflare, hCaptcha) fail at this layer. The browser tool is the answer.

Why this stack

A plain fetch with a tag stripper covers most public pages and costs nothing. Reaching for a headless browser on every URL would burn CPU and waste seconds when a 50 ms GET would do.

See also