Cloudflare has launched Kitesurf, a cloud-hosted browser built to help AI agents use websites and complete online tasks. The beta runs on Cloudflare Workers and is free through Browser Run, subject to account limits.
Developers can use Kitesurf for screenshots, HTML extraction, form completion, and other automated browser work. It supports existing tools through the Chrome DevTools Protocol, including Playwright, Puppeteer, and compatible AI clients.
Unlike a consumer browser, Kitesurf has no desktop interface for daily web use. Developers call it through APIs and automation libraries inside Workers. This setup lets software open pages, read content, capture files, and complete forms without maintaining a separate browser service or dedicated server infrastructure for each agent task.
Traditional browsers support tabs, extensions, themes, video, and detailed visual rendering for human users. AI agents often need structured page content, forms, links, and basic images instead. Cloudflare designed Kitesurf around those narrower requirements.
Cloudflare’s benchmark compared five median Browser Run tests across 14 web addresses. Kitesurf used 380 milliseconds of CPU time for screenshots, compared with Chromium’s 1,173 milliseconds. HTML extraction used 229 milliseconds, while Chromium used 877 milliseconds.
Memory use also fell during the same tests. Kitesurf used 57.8 MiB for screenshots against Chromium’s 271 MiB. For HTML extraction, it used 39.4 MiB, compared with 273.7 MiB for Chromium.
However, lower resource use came with slower completion times. Kitesurf took 1.8 times longer for screenshots and 1.7 times longer for HTML extraction. Cloudflare described the product as “significantly more efficient,” though its own data confirms the speed trade-off. Cloudflare published the full benchmark results in its launch announcement.
Cloudflare built Kitesurf around the idea that every page load may contain unsafe code. Each browser session starts fresh, while separate components receive access only to resources required for their tasks. This structure aims to limit data leaks between pages.
A dedicated SandboxOutbound worker handles requests for images, scripts, fonts, and stylesheets. It checks cross-origin rules, filters responses, adds browser headers, and keeps cookies in separate jars. Requests that fail its rules receive a 403 response.
Kitesurf divides browser work among an Engine, PageScript, and PageRenderer. The Engine stores session data and handles public connections. PageScript processes HTML, CSS, JavaScript, and WebAssembly, while PageRenderer creates screenshots and PDF output.
Cloudflare used Blitz modules, Firefox’s Stylo CSS parser, and the Boa JavaScript engine. The company traced its first working model to Obscura, an open-source Rust engine designed for browser automation.
Kitesurf currently renders TodoMVC applications, Wikipedia, Hacker News, Cloudflare’s blog, and much of its dashboard. Developers can access it through Browser Run Quick Actions or add a Kitesurf parameter to a CDP endpoint.
The browser passes more than 235,000 Web Platform Test subtests, according to updated Cloudflare documentation. The launch announcement initially cited more than 215,000 tests. These tests check standards support but do not confirm compatibility with every website.
Kitesurf does not yet support video playback, WebGL, long authenticated sessions, or some bot-check systems. Developers needing those functions can continue using Browser Run’s Chromium option.
Cloudflare built the browser in 12 weeks and released it while development continues. The company plans broader protocol coverage, better screenshots, more standards support, and lower resource use. It also plans to release Kitesurf as open-source software once the project is ready.
Also Read: Cloudflare to Layoff 1,100 Employees as CEO Says Job Cuts are Linked to AI, Not Weak Business