Blog ·

How to let your agent read any web page as clean markdown

Raw HTML wastes an agent’s context on markup. Here is how to give your agent a fetch tool that returns the page as clean markdown, on one prepaid key.

The problem

When an agent fetches a page itself, it gets raw HTML: navigation, scripts and markup that crowd out the content and burn context. Scraping services fix that, but each wants its own account and key.

With apifare

The scrape capability takes a URL and returns transformed markdown, never raw HTML. It needs no key of your own; its per-call price is listed on /capabilities.

How to set it up

  1. Get a key at /start and connect your agent with its setup guide.
  2. Ask. "Fetch https://example.com/pricing and summarize it." Your agent calls the capability with the URL and reads the markdown.

Pair it with search

Search finds the pages; fetch reads them. Together they give your agent a research loop on one balance, under the same caps, kill switch and receipts. See live web search.

Give this to your agent

set up https://apifare.com/SKILL.md — and it takes it from there: key, config, first call. $5 of calls free.

No agent open right now? Two manual steps: ① get a key (email in, key out) · ② give it to your agent — one config line, every client has a setup guide.

More from the blog

How to give your AI agent live web search without a search API key · How to pull Google Maps places and reviews into your AI agent · How to pull SEC filings and UK company records into your AI agent · How to get YouTube transcripts into your AI agent