Moonshot AI just released Kimi WebBridge, a browser extension that gives AI agents the ability to click, scroll, fill forms, and navigate websites,all while keeping your data on your machine instead of sending it through cloud servers. The extension works with Chrome and Edge browsers and integrates with multiple AI tools, including Claude Code, Cursor, and Kimi's own agent platform. For anyone concerned about privacy when automating browser tasks, this represents a meaningful shift in how AI agents can operate.

Most AI browser automation tools today route your activity through cloud infrastructure, which means your logged-in email accounts, banking sessions, and company internal tools all pass through a third party's servers. Kimi WebBridge takes a different approach. It pairs a local background service running on your computer with the browser extension, using Chrome DevTools Protocol,the same low-level interface that software developers use for debugging. Everything stays on your machine. Your bank account, your email, your company's internal tools: the agent can interact with all of them without Moonshot ever seeing the content.

What Can Kimi WebBridge Actually Do?

The practical applications are straightforward but time-saving. You could ask your AI agent to browse Amazon for mechanical keyboards under $150 with at least 4.5-star ratings and return a ranked comparison. The agent would understand your instruction and search the Amazon site visually instead of relying on API calls. You could ask it to scan LinkedIn job listings across multiple searches and compile them into a spreadsheet, or check prices for the same product across 10 retailers and report back the best deal. Any task that involves clicking through a website repeatedly,the kind of work that takes 20 minutes of manual effort,becomes a one-sentence prompt.

The extension officially supports Kimi Code CLI, Claude Code, Cursor, Codex, and Hermes, so it's not locked to the Kimi ecosystem. If you already use another AI coding or agent tool, WebBridge plugs into it.

How to Set Up Kimi WebBridge in Three Steps

• Download the Desktop App: You'll need Kimi Claw Desktop, which runs the local service that powers the extension. Mac users can download directly from the setup page at kimi.com/features/webbridge, while Windows users can install via PowerShell with a single command.
• Install the Browser Extension: Add the extension from the Chrome Web Store or manually through the setup page. Once installed, it connects to your local Kimi Claw service.
• Deploy Your First Agent: Open Kimi Claw Desktop, find Kimi Claw in the left sidebar, add a new agent, and select "On my computer" to run it locally. Then send a prompt like "Browse Amazon for a mechanical keyboard under $150 with 4.5+ stars" and it goes to work.

For users of other AI agents like Claude Code or Cursor, the setup page provides a connection command you paste into your agent, which connects it to the WebBridge service automatically. If the extension shows as disconnected, resending the connection command and restarting the app typically resolves the issue.

Why Moonshot AI's Kimi Model Matters for This Tool

WebBridge is powered by Moonshot AI's Kimi model family, built in China, and the underlying AI is more capable than many Western users realize. Kimi K2 launched in July 2025 as a 1-trillion-parameter open-source mixture-of-experts model, ranking first among open-source models and fifth overall on the LMSYS Arena leaderboard,a widely respected benchmark that compares AI models across reasoning, coding, and general knowledge tasks.

The latest version, K2.6, released in April 2026, now scores 58.6% on SWE-Bench Pro, a benchmark measuring real-world software engineering performance on actual GitHub issues. That puts it ahead of GPT-5.4 at 57.7% and Claude Opus 4.6 at 53.4%, demonstrating that Kimi is competitive with the most advanced models from OpenAI and Anthropic. Moonshot says K2.6 supports up to 300 parallel sub-agents executing across 4,000 coordinated steps simultaneously, the architecture WebBridge taps into when handling complex, multi-step browser tasks.

The Cursor Controversy That Put Kimi on the Map

If you haven't heard of Kimi before, there's a reason it landed on your radar in 2026: the Cursor controversy. On March 19, the $50 billion coding AI startup Cursor launched its Composer 2 model, marketing it as "frontier-level proprietary coding intelligence" built through "continued pretraining" and reinforcement learning. The announcement lasted less than 24 hours before a developer named Fynn intercepted API traffic and found a model identifier: kimi-k2p5-rl-0317-s515-fast. Elon Musk posted three words: "Yeah, it's Kimi 2.5." Moonshot's head of pretraining ran a tokenizer analysis and confirmed an identical match.

Cursor VP of Developer Education Lee Robinson acknowledged the open-source base within hours, insisting roughly 75% of the compute went into Cursor's own training pipeline. Cursor co-founder Aman Sanger called the omission "a miss from the start." Moonshot, for its part, took the high road,officially congratulating Cursor. The incident revealed that Kimi's open-source foundation had become valuable enough for a major AI startup to build on, and that Moonshot was willing to let competitors use its work.

How Does Kimi WebBridge Compare to Other Browser Automation Tools?

Browser automation for AI agents is getting crowded. Anthropic's computer use feature lets Claude interact with desktops. OpenAI's Operator and ChatGPT Atlas do similar things through a hosted cloud service. Google has DeepMind-powered agent experiments and Perplexity has its Comet Browser. The difference with Kimi WebBridge, at least for now, is the local-first architecture. Cloud-based browser agents are convenient but require routing your browsing activity through a third party. For anything involving personal accounts or sensitive data, that's a real consideration.

This privacy-first approach aligns with a broader trend in AI infrastructure. CoreWeave, a major AI cloud provider, recently launched CoreWeave Sandboxes, an execution layer that gives AI researchers and platform teams secure, isolated environments for running reinforcement learning, agent tool use, and model evaluation. The service runs directly within a customer's existing infrastructure, allowing teams to run complex AI workflows without adding a separate execution stack or routing sensitive data through third-party systems.

"CoreWeave Sandboxes solves a real gap in our AI research stack: secure, isolated code execution at scale directly in our existing compute," said Brian Belgodere, Senior Technical Staff Member, AI/ML Systems at IBM Research.

Brian Belgodere, Senior Technical Staff Member, AI/ML Systems at IBM Research

The pattern is clear: as AI agents move from experimental tools to production systems, organizations want to keep sensitive data and execution environments under their own control. Kimi WebBridge fits into this shift by letting you run browser automation locally, and CoreWeave Sandboxes extends the same principle to reinforcement learning and model evaluation workflows.

What Does This Mean for AI Adoption in Enterprise?

The release of Kimi WebBridge reflects a maturing AI market where privacy, control, and integration with existing tools matter as much as raw capability. Early AI agents were cloud-first because that's where the compute lived. Now that models are more efficient and local execution is feasible, companies are asking for alternatives that don't require sending every interaction to a third party.

Moonshot AI itself is in a strong financial position to support this shift. The company raised $2 billion at a $20 billion valuation in a Meituan-led round, nearly 5x its late-2025 valuation, as Kimi K2.6 crossed $200 million in annual recurring revenue. That funding gives Moonshot resources to compete globally while maintaining the open-source foundation that made Kimi attractive to companies like Cursor in the first place.

For developers and teams considering browser automation, the choice now comes down to priorities. If you need maximum convenience and don't mind cloud routing, OpenAI's Operator or Anthropic's computer use feature are solid options. If you're handling sensitive data or want to keep everything local, Kimi WebBridge offers a meaningful alternative. The fact that it works with multiple AI agents,not just Kimi's own tools,makes it a practical choice for teams already invested in other platforms.