Claude Code Chrome Extension: Setup, Uses and Fixes
Claude Code Chrome extension setup: connect Claude in Chrome with claude --chrome, test web apps, read console logs, fill forms and fix connection errors.
Josh Pocock8 min read
AI Coding
Claude in Chrome
Stride AI Academy
The Claude Code Chrome extension is the Claude in Chrome browser extension, connected to Claude Code so Claude can drive your real browser from the terminal or VS Code. Install the extension, then start Claude Code with claude --chrome. Claude can then open tabs, click, type, read console errors and fill forms while it builds your code, using the logins you already have.
This guide covers what the integration does, what you need, how to connect it, prompts that work, permissions, safety and the fixes for common errors. Everything here was checked against Claude Code's Chrome docs, the VS Code docs, and the support articles Getting started with Claude in Chrome, Claude in Chrome permissions guide and Use Claude in Chrome safely on November 18, 2026.
What Claude Code can do in Chrome
With Chrome connected, Claude can build a change and then check it in the same session. The docs list these uses:
- Live debugging. Read console errors and DOM state, then fix the code that caused them.
- Design checks. Build a UI from a mock, then open it to compare.
- Web app testing. Test form validation, check for visual regressions or walk through a user flow.
- Logged-in apps. Work in Google Docs, Gmail, Notion or any site you're signed into, without an API connector.
- Data extraction. Pull structured data from pages and save it locally.
- Form filling and repetitive tasks. Data entry across one or more sites.
- File uploads. Attach files from your machine to upload fields.
- Recordings. Save browser sessions as GIFs, or screenshots to disk.
Browser actions run in a visible Chrome window, so you can watch. Claude opens its own tabs and groups them into a Chrome tab group for the session. When it hits a login page or CAPTCHA, it stops and asks you to handle it.
What you need
| Requirement | Detail |
|---|---|
| Browser | Google Chrome or Microsoft Edge. Claude Code also connects to other Chromium browsers such as Brave, Arc, Vivaldi and Opera |
| Extension | Claude in Chrome version 1.0.36 or later |
| Claude Code | Installed and up to date. See how to install Claude Code |
| Plan | A direct Anthropic plan: Pro, Max, Ultra, Team or Enterprise |
| Sign-in | Claude Code signed in with /login |
Three things rule it out:
- API keys and long-lived tokens. If Claude Code authenticates with an API key or a token from
claude setup-token, Chrome integration stays off, even with--chrome. - Third-party providers. It isn't available through Amazon Bedrock, Google Cloud or Microsoft Foundry. You'd need a separate claude.ai account.
- WSL. Chrome integration isn't supported in Windows Subsystem for Linux.
On Enterprise, organisations with HIPAA enabled have it off by default until an Owner turns it on, and organisations with Zero Data Retention for Claude Code can't use it in Claude Code at all.
How to connect Claude Code to Chrome
Step 1: install and sign in to the extension
Add Claude in Chrome from the Chrome Web Store, sign in with your Claude account, and pin it from the puzzle-piece menu. Sign in to the same claude.ai account and organisation that Claude Code uses, or the connection fails.
Step 2: start Claude Code with the flag
claude --chromeThe first time, Claude Code shows a one-time dialog explaining the integration and site permissions. Press Enter to continue. Claude Code also writes a native messaging host file that Chrome reads on startup, so if the extension isn't detected the first time, restart Chrome.
Step 3: check the connection
Run this inside Claude Code:
/chromeIt's working when the panel shows "Status: Enabled" and "Extension: Installed". The same command lets you manage permissions, reconnect the extension and pick which browser to use if more than one is connected.
Turn it on by default
To skip the flag, run /chrome and choose the option that turns Chrome on by default. In VS Code, type /chrome in the prompt box and turn on Enabled by default. The CLI and the extension share this setting.
The trade-off: with it on, the browser tools and their instructions load into every session, which uses more context. If you care about context use, leave it off and pass --chrome only when you need the browser.
A project can't switch it on for you. If a repository's .claude/settings.json sets CLAUDE_CODE_ENABLE_CFC, Claude Code ignores it and shows a warning, so a repo you clone can't connect Claude to your browser.
In VS Code
Type @browser followed by a task in the Claude Code panel:
@browser go to localhost:3000 and check the console for errorsMore on the panel in how to use Claude Code in VS Code.
Example prompts
These come from the docs and are a good starting point.
Test a form you just changed:
I just updated the login form validation. Can you open localhost:3000,
try submitting the form with invalid data, and check if the error
messages appear correctly?Read console errors. Name what you're looking for, because logs get long:
Open the dashboard page and check the console for any errors when
the page loads.Extract data to a file:
Go to the product listings page and extract the name, price, and
availability for each item. Save the results as a CSV file.Record a demo:
Record a GIF showing how to complete the checkout flow, from adding
an item to the cart through to the confirmation page.To see every browser tool, run /mcp, select claude-in-chrome, then View tools. For scripted, headless browser testing in CI instead of your own Chrome window, see Playwright MCP with Claude Code.
Permissions and site access
There are two layers.
Claude Code's prompts. Before a browser action, Claude Code may ask. The prompt starts with "Claude in Chrome wants to" and includes an option to allow all actions on that site for the session. Your normal Claude Code permission settings still apply. In auto mode, when the classifier approves a browser call to a site, the extension skips its own per-site check for that call.
The extension's site permissions. Claude Code inherits site-level permissions from the extension. On a site that needs approval, you can choose "Allow this action", "Always allow actions on this site" or "Decline". To review or revoke sites, open the extension's side panel, click the three dots, choose "Extension settings" and go to Permissions, where you'll find "Your approved sites".
Even with "Always allow" on, the extension still asks before downloading files, entering sensitive information or granting authorizations.
File uploads follow your read rules: if a permission rule denies Read on a file, Claude can't upload it. Uploads are capped at 10 MB per upload, and Claude refuses files with credential-style names such as .env, .pem or .key files, or anything under .ssh.
Safety notes
Claude uses your real browser session, so it can reach anything you're signed into. Anthropic's safety article is direct that the risk from prompt injection, hidden instructions on a web page, is not zero, even with classifiers screening content and actions.
Practical rules from that article:
- Start with trusted sites. Avoid unknown sites and user-generated content for sensitive work.
- Use a separate browser profile with no banking, healthcare or government logins.
- Use "Manually approve" mode when you want to review every action.
- Write specific prompts, and stop the task if Claude visits unexpected sites or asks for sensitive information.
- Claude takes screenshots of active tabs, so keep sensitive pages closed.
GIF recordings capture everything visible, including account details on logged-in pages. Check them before sharing.
Troubleshooting
| Problem | Fix |
|---|---|
| "Browser extension is not connected" | Sign the extension in to the same claude.ai account and organisation as Claude Code, restart Chrome and Claude Code, then run /chrome to reconnect |
/chrome shows "Not detected" |
Install or enable the extension at chrome://extensions |
| "No tab available" | Ask Claude to open a new tab and try again |
| "Receiving end does not exist" | The extension went idle. Run /chrome and pick "Reconnect extension" |
| "No browsers are connected" | Open Chrome with the extension signed in, and check your network can reach bridge.claudeusercontent.com |
| Commands stop working mid-task | A JavaScript alert or confirm dialog may be blocking the page. Dismiss it and tell Claude to continue |
To check which organisation Claude Code uses, run /status and read the Organization row. On Windows, if you see a named pipe conflict (EADDRINUSE), close other Claude Code sessions using Chrome and restart. If the native messaging host crashes, reinstall Claude Code to regenerate it.
Learn to build with Claude Code
The Stride AI Coding course teaches Claude Code and Codex for building client tools, written for non-programmers, with weekly live calls to review your builds. Many of the free guides cover Claude Code setups, and the Claude Code community is where members compare theirs.
FAQ
Is there a Claude Code Chrome extension?
Claude Code uses the Claude in Chrome extension from the Chrome Web Store. Install it, sign in with your Claude account, then start Claude Code with claude --chrome to connect the two.
How do I enable Chrome in Claude Code?
Start Claude Code with claude --chrome, or run /chrome and turn it on by default. In the VS Code extension, type @browser with a task, or use /chrome to turn on Enabled by default.
Which plans can use Claude Code with Chrome?
You need a direct Anthropic plan: Pro, Max, Ultra, Team or Enterprise, with Claude Code signed in through /login. It doesn't work with API keys, claude setup-token tokens or third-party providers like Amazon Bedrock.
Does Claude Code work with Edge or Brave?
Claude Code's docs say the integration works with Google Chrome and Microsoft Edge, and that it also connects in Brave, Arc, Vivaldi and Opera. It isn't supported in WSL.
Why does Claude Code say the browser extension is not connected?
Usually the extension isn't running, or it's signed in to a different claude.ai account or organisation than Claude Code. Match the accounts, restart Chrome and Claude Code, then run /chrome and reconnect.
Is it safe to let Claude Code control my browser?
It acts with your logged-in sessions, so prompt injection is a real risk. Use trusted sites, a separate browser profile without financial or medical logins, and manual approval for anything sensitive.