How to Use Claude in Chrome with Claude Code

To use Claude in Chrome with Claude Code, install the browser extension, connect it to Claude Desktop, and let Claude Code handle the codebase while Claude in Chrome handles live browser testing. The two tools are designed to work together: Claude Code builds, Claude in Chrome tests and verifies — closing the loop between writing code and confirming it works in a real browser.

What Is Claude in Chrome?

Claude in Chrome is an official browser extension that lets Claude act as an AI agent directly inside Google Chrome. It opens as a persistent side panel while you browse, allowing Claude to read page content, click buttons, fill forms, and navigate websites on your behalf — not just answer questions about them.

The extension works by taking screenshots of your active tab and interacting with the page's underlying structure. You can ask Claude to handle tasks across multiple tabs, schedule recurring browser jobs, record a workflow once and have Claude repeat it, and connect it to Claude Desktop or Claude Code for end-to-end software development workflows.

Claude in Chrome is available on Max, Team, and Enterprise plans. Get started with Claude in Chrome using Anthropic's official setup guide.

How Do You Install and Set Up Claude in Chrome?

1. Open Google Chrome on a desktop computer (macOS or Windows). The extension does not work on mobile devices or other browsers.
2. Go to the Chrome Web Store and search for "Claude in Chrome." Click "Add to Chrome" to install the extension.
3. Pin the extension by clicking the puzzle piece icon in the Chrome toolbar, then clicking the thumbtack next to "Claude" for easy access.
4. Open the side panel by clicking the pinned Claude icon, then sign in with your Claude account credentials.
5. Grant the required permissions when prompted. These cover the chat interface, reading page text, simulating clicks and keystrokes, managing Claude's tabs separately from yours, and scheduled tasks.
6. Connect to Claude Desktop (optional but recommended for Claude Code workflows): Open the Claude Desktop app, click your initials in the lower-left corner, go to Settings, find the Connectors section, locate "Claude in Chrome," and toggle it on. This enables Claude Desktop to trigger browser tasks without switching windows.
7. Choose your operating mode in any new chat: "Ask before acting" (recommended) or "Act without asking." Note that the connector is disabled by default per conversation and must be enabled each time.

How Do You Use Claude in Chrome with Claude Code?

The most powerful workflow combines Claude Code in your terminal with Claude in Chrome in your browser. Here is how the build-test-verify loop works in practice:

1. Use Claude Code in your terminal to write or edit a feature in your codebase.
2. Claude Code builds and runs the feature, then triggers Claude in Chrome to load the staging site.
3. Claude in Chrome reads the rendered page, checks console errors and network responses from the page's underlying structure, and returns a diagnostic report to your terminal — all without you manually opening browser developer tools.
4. If issues are found, Claude Code applies fixes and the cycle repeats.

This end-to-end loop means you stay in your editor while Claude handles both the code and the browser verification. As the source material puts it, the two tools are designed to work together: "Claude Code builds, Claude in Chrome tests and verifies."

What Are the Most Useful Things You Can Do with Claude in Chrome?

Research and summarization across multiple tabs

Ask Claude to open several competitor websites, read the content of each, and synthesize a comparison table in the side panel — turning a lengthy manual research task into a single prompt.

Email drafting and inbox triage

A busy professional can ask Claude to scan their Gmail inbox, categorize messages, and draft responses — without reading every message individually.

Automated recurring data extraction

Record a workflow once using the "Teach Claude" feature, then schedule it to run automatically on a recurring basis without any further manual steps.

Form filling with approval checkpoints

Claude pre-fills form fields based on your standard information, then pauses for your review before submitting — reducing repetitive data entry to a single confirmation step.

Software development build-test-verify loop

As described above, Claude Code and Claude in Chrome together handle the full cycle from writing code to verifying it in a live browser environment.

When Should You Use Claude in Chrome vs. Other Tools?

Is Claude in Chrome Safe to Use?

Browser use is a beta feature with real security implications, and Anthropic is direct about this. Claude can see everything visible on your screen when the panel is open, and malicious content on web pages could potentially influence Claude's actions — a risk known as prompt injection.

Anthropic provides two operating modes: "Ask before acting" and "Act without asking." The safer, approval-based mode is recommended for most tasks, especially on unfamiliar or untrusted websites. Read Anthropic's full safety guidance for Claude in Chrome before running autonomous workflows.

Key safety practices

• Default to "Ask before acting." This makes Claude present a plan before taking any browser action, giving you a chance to catch unexpected steps.
• Navigate away from sensitive pages before opening the panel. When the side panel opens, Claude takes a screenshot of the active tab. Passwords, financial data, and medical records visible at that moment can be seen by Claude.
• Review domain permissions individually. The extension requires approval before running scripts on any new website. Never bulk-approve all domains — review each one, especially sites where you are logged into accounts.
• Stop immediately if Claude proposes an action you did not request. This is a signal of potential prompt injection. Report unexpected behavior to Anthropic's security team.
• Use a separate Chrome profile without access to sensitive accounts if you want to experiment freely without risk.

What Pitfalls Should You Avoid?

• Running long autonomous sessions without monitoring. Browser interactions consume more compute than regular chat, and long-running scheduled tasks compound this. Break large workflows into smaller approved steps.
• Recorded workflows breaking after a site redesign. The extension pauses and waits for your input when it encounters an unexpected page state — a new UI element, a CAPTCHA, or a login prompt. Resolve the blocker manually, then instruct Claude to resume. Re-record workflows periodically for sites that update frequently.
• Forgetting to enable the connector each conversation. The Claude in Chrome connector is disabled by default per conversation and must be enabled each time you start a new chat.

How Do You Get the Most Out of the Claude Code + Chrome Integration?

The build-test-verify loop is the flagship use case for combining these two tools. A practical approach:

• Keep Claude Code focused on the codebase — writing, editing, and running tests.
• Use Claude in Chrome only when you need to verify something in a real browser: rendered UI, console errors, network responses, or user-facing behavior that only appears after login or interaction.
• Connect Claude Desktop as the bridge between the two tools so Claude Code can trigger browser tasks without you switching windows.
• Use "Ask before acting" during development so you can review each browser step before it runs — especially useful when testing forms or authenticated pages.

For a deeper look at permissions and what each one allows, see the Claude in Chrome Permissions Guide.