Claude for Chrome Is Now Generally Available: What the Browser Extension Means for Your Dev Workflow
Anthropic's Claude for Chrome extension has exited preview and is now generally available to all direct paid plan users, bringing browser automation into Claude Code.
As of July 1, 2026, the Claude for Chrome extension has moved from preview to general availability for everyone on a direct Anthropic paid plan. Pro, Max, Team, and Enterprise users can now use it without waitlists or early-access caveats.
The extension connects Claude Code to your Chrome browser, and the practical upshot is straightforward: you no longer have to stop, switch to a browser, poke around manually, then switch back to your terminal. Claude handles that round-trip for you.
What the Extension Actually Does
When Claude for Chrome is connected to Claude Code, Claude can open tabs, click through pages, fill in forms, read your browser’s console output, and share your existing login session. That last point matters: Claude sees the sites you’re already signed into, so it can access your internal tools, staging environments, and anything else that sits behind authentication without you needing to hand over credentials separately.
The concrete workflow looks like this: build in your terminal with Claude Code, deploy to a URL, and then ask Claude to open it in Chrome and verify that the button placement matches the Figma spec, or that the form submits correctly, or that there are no console errors on load. Claude reads the DOM state, picks up network errors, and can fix issues in your code without you leaving the session.
Beyond basic testing, there are a few other capabilities worth knowing about:
- Session recording. Claude can record browser interactions as GIFs, useful for documenting bugs or sharing reproduction steps with a teammate.
- Workflow teaching. Record yourself doing a repetitive browser sequence and Claude learns to repeat it. Useful for things like weekly reporting workflows or QA checks that always follow the same pattern.
- Scheduled tasks. You can set recurring browser tasks. Ask Claude to check your site’s core pages every morning and flag regressions, then leave it to run.
The Technical Setup
You need Claude Code 2.0.73 or higher and extension version 1.0.36 or higher. It works with Chrome and Edge, but not Brave, Arc, or other Chromium-based browsers at this point.
To enable Chrome from the CLI, run /chrome inside a Claude Code session. If you want it on by default without passing --chrome every time, select “Enabled by default” from that same menu. In the VS Code extension, if the Chrome extension is installed, it’s available automatically with no extra flag.
One thing to be aware of: enabling Chrome by default increases context usage because the browser tools are always loaded into the session, even when you’re not actively using them. If you notice token consumption climbing faster than expected, it’s worth disabling the default and using --chrome only on sessions where you actually need it.
Internally, the extension runs on Claude Haiku 4.5 rather than Sonnet, keeping things fast and cost-efficient for multi-step browser interactions.
Permissions and Safety
Claude interacts with websites on your behalf, so the permissions model is worth understanding. Site-level access is controlled through the Chrome extension settings, where you can specify which sites Claude is allowed to browse and interact with. In plan mode, read-only browser actions (reading page content, checking console state) run without prompting. Actions that change state, like clicking a button or submitting a form, prompt for approval first.
Certain site categories are blocked by default: financial services, banking, cryptocurrency exchanges, adult content, and pirated content. Anthropic has also added safety classifiers to the extension as part of the GA release, though they still describe the feature as carrying some risk given the nature of direct browser interaction.
For Team and Enterprise accounts, admins can configure allowlists and blocklists to control which domains Claude can access across the organisation.
Who This Is For, and Who It Isn’t
The extension is available on all direct Anthropic paid plans. If you access Claude exclusively through Amazon Bedrock, Google Cloud’s Agent Platform, or Microsoft Foundry, this feature is not available through those routes. You would need a separate claude.ai account to use it.
For individual developers, the most immediate benefit is the reduction in context-switching during testing and debugging cycles. The time you spend copying a console error out of Chrome and pasting it into your terminal adds up. Having Claude read the error and propose a fix in the same session is a small thing that compounds over a working day.
For teams, the scheduling and workflow features open up some useful automation territory. Regression checks, form validation testing, and structured QA sequences that previously needed a person to babysit can run on a schedule without any manual involvement.
The GA designation also matters in an organisational sense. Preview features carry an implicit “handle with care” label. GA means Anthropic considers it stable enough to deploy broadly, which makes it easier to justify rolling it out across an engineering team rather than leaving it to individual early adopters.
The Bigger Picture
Claude Code started as a terminal-first tool. The Chrome extension, combined with the subagent background processing and VS Code integration that have also shipped this year, suggests the direction: an AI coding assistant that follows you across your actual working environment rather than waiting for you to bring work to it.
The Claude for Chrome extension is available on the Chrome Web Store, and setup documentation is in the Claude Code Chrome integration docs.