PagePith returns readable Markdown and HTML from the pages your product depends on — including JavaScript apps, PDFs, and most bot-protected sites. One endpoint chooses the right retrieval path automatically, so your code never has to.
Use plain HTTPS from any language. The response always includes the final URL, title, readable Markdown, HTML, fetch time, content length, and cache status.
curl -X POST https://api.pagepith.com/v1/api/scrape \
-H "Authorization: Bearer $PAGEPITH_KEY" \
-H "Content-Type: application/json" \
-d '{ "url": "https://example.com/pricing" }'{
"url": "https://example.com/pricing",
"finalUrl": "https://example.com/pricing",
"title": "Pricing · Acme",
"markdown": "# Pricing\n\n## Pro\n$79 / month…",
"html": "<html>…</html>",
"contentLength": 1842,
"fetchedAt": "2026-08-11T14:32:08.000Z",
"cached": false
}PagePith evaluates each page and escalates only when the simple path is not enough. Your integration stays the same whether the content came from HTML, hydration data, a browser render, a PDF, or the fallback tier.
PagePith tries a direct fetch first, checks metadata and hydration data when useful, renders JavaScript when necessary, and keeps a final fallback for difficult pages.
Article pages are distilled with Mozilla Readability. Other pages keep their useful sections while navigation, scripts, ads, cookie banners, and other noise are removed from the Markdown.
Use the synchronous endpoint when a caller can wait. For ingestion jobs, queue the same scrape with a callback URL and receive the result when it is ready. Queued jobs retry automatically.
Successful responses are cached for seven days by default. Bypass the cache per request with forceFresh, invalidate a URL explicitly, and branch on the cached flag in every response.
Create a monitor for any public URL. PagePith checks it on a schedule, stores the check history and line-level change summary, then notifies you by email or signed webhook.
PagePith ships a remote, streamable HTTP MCP server. Connect any compatible client and your agent gets a focused read_url tool backed by the same scraper — without a local browser or MCP process to run.
One dependable tool. read_url accepts a URL and an optional fresh flag, then returns the page title, final source URL, and Markdown.
Same difficult-page handling. JavaScript sites, text-layer PDFs, and most bot-protected pages use PagePith’s full retrieval stack.
Same key, same meter. MCP requests use your PagePith API key and are billed exactly once, only after a successful scrape.
{
"mcpServers": {
"pagepith": {
"url": "https://mcp.pagepith.com/mcp",
"headers": {
"Authorization": "Bearer $PAGEPITH_KEY"
}
}
}
}Every plan includes the scraping API, monitoring, and hosted MCP server. One successful page is one credit; failed scrapes are not billed.
Effective rate on the recommended plan: $5.27 per 1,000 pages.