Scrape and monitor websites

Map a site's URLs, crawl or extract its pages as clean markdown, and get told when a page changes, through one MCP connection with several scraping providers behind it.

The problem

Your agent needs the content of web pages, not search snippets, and needs to notice when a pricing or docs page changes. Running your own scraper means proxies and headless browsers, and each scraping API wants its own account and plan.

Steps

5
  1. Discover all the URLs on a website$0.000143 per site map with links

    List the URLs on a site, with titles and descriptions, to pick the pages worth reading. Open web & SEO data · web.map

  2. Crawl a website and extract its pages$0.000143 per page

    Crawl the site to a set depth and get each page as markdown. Open web & SEO data · web.crawl

  3. Extract clean text from a web page$0.000143 per page

    Extract the clean text of one page, with menus and clutter removed. Open web & SEO data · web.extract

  4. Start a scheduled web-page change monitorNo charge

    Start a scheduled check on a page, at most once an hour, that records each change. Open web & SEO data · web.monitor.create

  5. Read new changes from page monitorsNo charge

    Read the latest changes detected on your monitored pages. Open web & SEO data · web.monitor.changes

Example cost for 100 rows

  • Discover all the URLs on a website$0.0143

    100 x $0.000143 per site map with links

  • Crawl a website and extract its pages$0.0143

    100 x $0.000143 per page

  • Extract clean text from a web page$0.0143

    100 x $0.000143 per page

  • Start a scheduled web-page change monitorNo charge

    100 x No charge

  • Read new changes from page monitorsNo charge

    100 x No charge

At most, for 100 rows through every step$0.0429

Steps charged only when a result is found cost less when nothing is found. Paid from your prepaid balance at the quoted customer price. No top-up fee. Pay per call; each call’s price is shown before it runs. Each run is quoted before it starts.

Try it

Paste this into your agent

Start free, no card needed. Your agent runs each step and reports the cost.

Use Ahel. Map the URLs on competitor.com with web.map and pick the pricing, product and docs pages. Extract each one with web.extract and summarize what they offer. Then start a monitor on the pricing page with web.monitor.create, checking once a day, and next time read what changed with web.monitor.changes. Tell me what it cost.

Which scrapers run it

Page extraction can run on Ahel's self-hosted engines or on hosted providers such as Firecrawl. Ahel picks by measured success and moves to the next provider when one fails; each task page lists them in order with their prices. You do not need an account with any of them.

What monitoring costs

Starting a monitor and reading its changes have no charge. Each successful page check uses a small amount of your balance, shown on the monitor's task page, and failed checks are free.

Public pages only

Monitors watch public pages whose terms allow it, respect robots.txt and never get around logins or access controls.

Set up Ahel in your agent

  • Claude Code

    Run once, then approve the sign-in: claude mcp add --transport http --scope user ahel 'https://mcp.ahel.ai'

  • Codex

    Run once, then approve the sign-in: codex mcp add ahel --url 'https://mcp.ahel.ai' && codex mcp login ahel

  • Claude.ai

    Add a custom connector with the server address https://mcp.ahel.ai and approve the sign-in.