Webfetch, a web scraping API built and run by Tuxxin, is open at webfetch.io. A request names a URL and gets the page back as rendered HTML, plain text, markdown, or JSON pulled out with CSS or XPath selectors. The technical announcement, with the full credit table and every plan, is on the Webfetch blog.
Compatible with ScrapingBee
Webfetch takes the parameters, defaults and credit costs of ScrapingBee's HTML API. Code written for ScrapingBee moves over by changing the base URL and the key, unless it depends on AI extraction, scraping_config, ScrapingBee's dedicated APIs, or screenshots returned as the response body. Response headers change prefix from Spb- to Wf-, so code that reads them needs the new prefix. Plans with the same credit allowance cost 38 to 40% less, based on ScrapingBee's published prices on October 8, 2026.
Plans and the free tier
Signup confirms a free account with a six-digit code sent to its email address and asks for no card. Free credits start at 1,000, and the 1st of each month adds 250, up to a 1,000 balance. Self-serve plans run from Starter at $11.99 a month for 75,000 credits to Business+ at $359.99 a month for 8,000,000, and Custom plans on dedicated hardware start at $499 a month. Each subscription keeps the monthly price it started at for as long as it stays active.
Until December 7, 2026, the first 500 new accounts to enter LAUNCH500 at signup get 500 one-time bonus credits.
What a request gets
- Webfetch bills a request only when the target's status is 2xx, 404, 410 or 413. Blocked requests, timeouts, other errors and anti-bot challenge pages cost nothing, even when a challenge page arrives with HTTP 200.
- With
mode=auto, Webfetch starts with the cheapest configuration and moves up after a block, a challenge page, a timeout, a failed connection or render, or a server error, charging for the one attempt that worked. - Screenshots are captured on Tuxxin's own servers, with webshot.site, Tuxxin's screenshot API, as the fallback.
- robots.txt is honored by default, and private and internal network addresses are refused on every request and every redirect.
For developers and AI agents
Official SDKs for JavaScript, Python and PHP are on npm, PyPI and Packagist, with the source on GitHub. Webfetch also runs an MCP server at https://webfetch.io/mcp, so an AI client can fetch a page as markdown or extract structured data with the same API key.
Part of the Tuxxin Stack
Webfetch joins the other services Tuxxin builds and runs, including webshot.site for screenshots, whack.sh for URL threat scanning, issued.live for certificate and DNS data, and WorldIP.io for IP intelligence. The projects page lists all of them, Webfetch included.
Accounts open at webfetch.io/signup, and the home page runs a real scrape without a key.