Cloudflare has launched Kitesurf, a new browser designed specifically for artificial intelligence agents rather than human users. The company says the browser is built to make it easier, faster and more cost-efficient for AI agents to navigate websites, extract information and perform web-based tasks.
Unlike traditional browsers such as Chrome, Kitesurf is designed around what AI agents actually need when interacting with the internet. Cloudflare describes it as an “agent-first” browser that can run in lightweight V8 isolates on its Workers platform.
What Is Cloudflare Kitesurf?
Kitesurf is a cloud-hosted, stateless browser engine intended to be used by AI agents.
AI agents increasingly need to interact with websites rather than simply read information supplied through APIs. They may need to open pages, extract content, take screenshots, generate PDFs or perform other browser-based tasks.
Traditional browsers were built primarily for humans and therefore include many features that an AI agent may not need.
Cloudflare’s approach is to strip away much of that overhead and provide a browser environment specifically optimised for machine-driven tasks.
Why Cloudflare Built a Browser for AI
AI agents are rapidly moving beyond simple question-and-answer interactions.
Modern agents can browse websites, research information, fill out forms, interact with online services and execute multi-step workflows.
However, running a conventional browser for every AI task can be expensive in terms of CPU and memory.
Cloudflare says Kitesurf is significantly more efficient than Chromium for common agentic tasks such as screenshots and HTML extraction.
That could become increasingly important as companies deploy thousands or millions of AI agents simultaneously.
Kitesurf Runs on Cloudflare Workers
One of the key technical features of Kitesurf is that it runs inside V8 isolates on Cloudflare Workers.
Instead of requiring a traditional browser installation on a local computer or dedicated server, Kitesurf can operate within Cloudflare’s distributed cloud infrastructure.
The architecture is designed to make browser sessions lightweight, isolated and highly scalable.
Cloudflare says Kitesurf was built using Rust and WebAssembly technologies and is designed around ephemeral workloads — meaning a browser environment can exist for the duration of an agent’s task rather than remaining active continuously.
How Is Kitesurf Different From Chrome?
The biggest difference is the intended user.
Chrome and other mainstream browsers are designed for people. They need to support everything from video playback and extensions to complex visual interfaces and long-running sessions.
Kitesurf instead focuses on tasks that AI agents commonly perform.
These include:
- Navigating web pages
- Extracting HTML and content
- Taking screenshots
- Generating PDFs
- Automating browser workflows
- Performing short-lived web tasks
- Scaling browser sessions for large numbers of agents
Cloudflare says Kitesurf can use substantially fewer resources than Chromium for these kinds of workloads.
Why Lower Resource Usage Matters
Running a browser for one AI agent may not seem particularly expensive.
The economics change dramatically when an organisation operates thousands of agents.
If every agent requires a full Chromium environment, memory and CPU consumption can quickly become a major infrastructure expense.
A lightweight browser could reduce the cost of running those workloads and make large-scale autonomous web activity more practical.
This is one of the central ideas behind Kitesurf: AI agents should not necessarily need a human-sized browser to use the web.
Kitesurf Is Part of Cloudflare’s Bigger AI Strategy
Kitesurf is not an isolated Cloudflare project.
The company has been building a broader infrastructure stack aimed at AI agents, including Browser Run, which allows agents to operate browser sessions on Cloudflare’s global network. Cloudflare says Browser Run can be used for tasks such as navigating websites, reading pages, filling forms, extracting information and taking screenshots.
Cloudflare has also introduced tools for agent authentication, sandboxes and other infrastructure designed to help developers deploy AI agents.
Kitesurf therefore represents another step in Cloudflare’s attempt to provide infrastructure for an internet increasingly accessed by autonomous software.
Could Kitesurf Change How AI Agents Use the Web?
Potentially.
Today’s AI agents often rely on a mixture of APIs, web scraping systems and conventional browsers.
But APIs are not available for every website, while conventional browsers can be resource-intensive.
An agent-focused browser could provide another layer between AI models and the web.
Instead of designing a browser around human interaction, developers can build applications around the specific needs of autonomous agents.
This could make browser-based AI automation more scalable.
What About Websites and Bot Protection?
The rise of AI browsers also creates new challenges for website operators.
Websites have traditionally distinguished between human visitors, search crawlers and automated bots. AI agents introduce a new category: software that can potentially behave like a user while performing tasks on behalf of a person or business.
That raises questions around authentication, permissions, rate limits, content access and payment.
Cloudflare itself has been working on what it calls an agent-first web, including tools to help website owners understand whether their sites are ready for AI agents.
The company has also developed systems for giving agents structured responses instead of traditional HTML error pages, reducing the amount of unnecessary information an AI model has to process.
Security Will Be Critical
Giving AI agents the ability to browse the internet also creates security risks.
An agent may encounter malicious instructions embedded within webpages, phishing content or prompt-injection attacks designed to manipulate its behaviour.
Because an AI agent can potentially access accounts, submit forms or perform actions, a compromised browsing session could have consequences beyond simply displaying an unsafe webpage.
For that reason, isolation and strict control over what an agent can access will be critical as agentic browsers become more widely adopted.
Cloudflare’s use of isolated execution environments is intended to help address some of these infrastructure concerns.
Kitesurf Is Not a Replacement for Chrome
Despite the attention surrounding the launch, Kitesurf is not intended to replace consumer browsers.
It is primarily infrastructure for developers building AI agents and automated applications.
People will continue to use conventional browsers for tasks that require rich interfaces, multimedia, extensions and long-running interactive sessions.
Kitesurf instead targets a different market: software that needs to use the web autonomously.
What Happens Next?
Cloudflare says Kitesurf is currently available in beta through its Browser Run offering, with plans to open-source the browser.
If the technology proves reliable at scale, it could become an important component of the infrastructure supporting AI agents.
The larger trend is perhaps even more significant than Kitesurf itself.
The web was originally designed around humans using browsers. With AI agents increasingly acting on behalf of humans, companies are now beginning to build infrastructure specifically for machines to navigate and interact with the internet.
Bottom Line
Cloudflare’s Kitesurf represents a significant shift in browser technology: a browser designed not for people, but for AI agents.
By focusing on lightweight rendering, isolation and scalability, Cloudflare hopes to make browser-based AI tasks considerably more efficient than running conventional browsers.
The launch also signals where the next phase of the internet could be heading. As AI agents become capable of researching, shopping, filling forms and completing complex digital tasks, the browser itself may need to evolve from a tool for human interaction into infrastructure designed for autonomous software.
