HomeNews

Cloudflare Open-Sources Kitesurf Browser for AI Agents

Kyle Morrison
Kyle Morrison
•
Nov 14, 2026 10:39 PM
•
0
 min read
Select Emergent as your Preferred news source
Cloudflare Open-Sources Kitesurf Browser for AI Agents

💡 TL;DR

  • Cloudflare published Kitesurf browser engine source code on GitHub under Apache 2.0 on November 13, 2026.
  • The stack combines Blitz, Stylo, and Boa in WebAssembly for lower CPU and memory footprint than Chromium.
  • Developers can now self-host and modify the engine, though a Workers Paid plan is still required.

Cloudflare published the source code for Kitesurf, its stateless browser engine designed for AI agents, on GitHub under the Apache 2.0 license on November 13, 2026. The release lets developers inspect, modify, and self-host the engine rather than relying solely on Cloudflare's managed Browser Run APIs. Engineer Celso Martinho describes the project as a deliberate alternative to Chromium, optimized for lower CPU and memory overhead while scaling across Cloudflare's global Workers network.

Release Date and Availability

Officially released on November 13, 2026, Kitesurf is now available on GitHub for developers who want to run the engine locally or deploy custom instances to Cloudflare Workers. The engine was first introduced in August 2026 and extended in September with WebMCP support. While the code is open, self-hosting still requires a Workers Paid plan because Kitesurf relies on Dynamic Workers to create isolated execution contexts for each page.

Architecture and Technical Stack

Kitesurf assembles three existing Rust projects into a cohesive browser runtime compiled to WebAssembly. Blitz handles HTML parsing, the DOM, layout, and painting. Stylo manages CSS parsing, cascading, and computed styles. Boa executes JavaScript inside WebAssembly as a fallback when JavaScript cannot run in Dynamic Workers. The engine speaks the Chrome DevTools Protocol, so existing Puppeteer scripts connect without rewrites.

Martinho's launch post frames the architecture as a trade-off: "Our main goal is to give AI agents a lightweight, efficient browser, one that can load websites, execute JavaScript, and interact with pages while using less CPU and memory than a full browser like Chromium." The team has not published specific CPU or memory benchmarks but points readers to a landing page for Web Platform Tests results.

Use Cases for AI Agents

The stateless design targets scenarios where agents need to interact with web pages at scale without the overhead of persistent browser sessions. Common workflows include web scraping, form submission, content extraction, and automated testing. Because Kitesurf runs on Cloudflare Workers, it scales on demand across Cloudflare's edge network without managing server infrastructure. The engine's compatibility with the Chrome DevTools Protocol means developers can reuse existing automation scripts built for Puppeteer or Playwright.

The Apache 2.0 license removes friction for teams that need to audit, extend, or embed the engine in proprietary systems. Self-hosting also gives organizations control over data residency and compliance requirements that managed APIs cannot always satisfy.

What This Means

Cloudflare's decision to open-source Kitesurf reflects a broader industry shift toward transparent, modular agent infrastructure. By publishing the engine under a permissive license, Cloudflare invites contributions while retaining monetization through the Workers platform. Teams building AI agents now have a lightweight alternative to Chromium-based solutions, though the Workers Paid plan requirement limits the audience to commercial deployments. The release positions Cloudflare as a key player in the emerging category of agent-native browser engines.

About the writer

Kyle Morrison leads Community at Emergent, where he brings together founders, operators, and SMB owners who are building real software with AI agents. He focuses on helping builders go from their first app to production use.

Start Building
on Emergent today
Try Emergent