Anthropic has rolled out a research preview of its new Claude for Chrome agent, giving select subscribers access to an AI assistant directly inside Google Chrome. The launch marks a major step in bringing artificial intelligence into the world’s most popular browser.
The feature is initially available to 1,000 Max plan customers, who pay between $100 and $200 per month. A waitlist has been opened for other users interested in trying out the extension.
With Claude for Chrome, users can chat with the AI in a persistent side window that keeps track of their browsing context. They can also authorize Claude to perform tasks directly in the browser, such as filling out forms or completing actions on websites. This move places Anthropic in direct competition with other tech companies racing to embed AI agents into everyday internet use.
The browser is quickly becoming the next big battleground for AI. Perplexity recently launched its Comet browser with a built-in assistant, while OpenAI is rumored to be preparing its own AI-powered browser. Google has also begun rolling out Gemini integrations with Chrome, even as it faces antitrust scrutiny that could force it to sell the browser. Reports suggest Perplexity has floated a $34.5 billion bid for Chrome, and OpenAI CEO Sam Altman has hinted at interest in buying it too.
With AI agents gaining new powers inside browsers, safety is emerging as a major concern. One of the biggest risks is prompt injection, where malicious websites embed hidden instructions to trick an AI into actions users never intended. Brave’s security team recently flagged such vulnerabilities in Perplexity’s Comet, though the company says they have been fixed.
Anthropic says it has taken steps to protect users, claiming its safeguards reduced the success rate of these attacks from 23.6% to 11.2% in internal testing. The company acknowledges, however, that risks remain as AI technology spreads deeper into web browsing.
For now, Claude for Chrome is available only to a small group of testers. But if the experiment proves successful, it could become one of the most widely used AI tools in daily online life, reshaping how people browse, chat, and get tasks done on the internet.
