Anthropic has introduced a native browser inside its Claude Cowork AI assistant, enabling users to access the web within the app itself. This new feature is enabled by default for Enterprise, Pro, Max, and Team subscribers, offering greater integration and streamlined access to web data without launching an external browser or extension.
- Claude Cowork now includes a native side-panel browser.
- Feature is on by default for most paid user tiers.
- New browser avoids triggering EU Digital Markets Act choice screens.
What happened
Anthropic has integrated a built-in web browser directly into its Claude Cowork platform. When a user prompt requires internet access, Claude automatically opens this browser in a side panel. This replaces the prior reliance on a Chrome extension to provide browsing functionality. The built-in browser is switched on by default for Enterprise users and is being rolled out to Pro, Max, and Team subscribers, with an option to revert to the extension in settings.
The new browser is designed narrowly to fetch publicly available information and does not have access to personal account logins such as email or banking services unless entered manually. This approach makes the browser a lightweight, task-focused tool embedded inside Claude rather than a full web browser experience.
Why it matters
By embedding a browser directly into Claude Cowork, Anthropic reduces dependency on Google Chrome, which dominates roughly 70% of global web traffic. This move bypasses the Digital Markets Act (DMA) in Europe, which requires gatekeeper platforms to present choice screens to users selecting default browsers or search providers. Because Claude’s built-in browser is part of the app and not considered a core platform, it does not trigger DMA regulations or choice screen obligations.
While this offers convenience and regulatory circumvention benefits, it also raises security and privacy concerns. Risks like prompt injection attacks, where malicious web content could influence or hijack AI instructions, remain present. The solution reuses existing guardrails from the Chrome extension but is not immune to vulnerabilities, as evidenced by past reports of virtual machine escapes and credential exposure on Macs.
What to watch next
The adoption pace of Anthropic’s in-app browser among its user tiers and the reaction from regulators, especially in the EU, will be key to monitor. This approach may inspire other AI assistant providers to embed browsers internally, potentially reshaping how browser competition and user choice are regulated under new digital market laws.
Security developments will also be critical to follow as the implications of prompt injection and web-based AI exploits become clearer. Anthropic and competitors like OpenAI, which retired its standalone Atlas browser recently, face ongoing challenges to balance seamless web access within AI tools against safeguarding user data and systems from attack.