Chrome Browser

Chrome Agent Bridge

Use the chrome-browser CLI to control the user's actual Chrome session through the installed Chrome Agent Bridge extension.

Setup checklist

Before first use:

1. Install the Chrome extension from the repo or Chrome Web Store.
2. Build the Rust binaries.
3. Install the native messaging host manifest.
4. Verify with chrome-browser ping.

Core commands

chrome-browser ping
chrome-browser navigate '{"url":"https://example.com"}'
chrome-browser getContent '{"format":"annotated"}'
chrome-browser click '{"text":"Sign in"}'
chrome-browser type '{"selector":"input[name=q]","text":"hello","submit":true}'
chrome-browser fillForm '{"fields":[{"selector":"#email","value":"a@b.com"}]}'
chrome-browser screenshot '{}'
chrome-browser evaluate '{"script":"return document.title"}'
chrome-browser scroll '{"y":500}'
chrome-browser uploadFile '{"selector":"input[type=file]","path":"/tmp/file.pdf"}'

Recommended workflow

1. chrome-browser ping
2. chrome-browser navigate ...
3. chrome-browser getContent '{"format":"annotated"}'
4. Use selectors or visible text from the annotated output for click / type / fillForm
5. Use screenshot when visual confirmation matters

Notes

• This controls the real Chrome browser, not a headless copy.
• Authenticated sessions and cookies are preserved because actions happen inside the user's actual browser.
• Chrome Web Store pages themselves may restrict extension automation.
• evaluate runs in the page's main world and can return JSON-serializable values.

Supported actions

• ping
• navigate
• getContent
• getInteractables
• click
• type
• fillForm
• waitFor
• scroll
• evaluate
• screenshot
• listTabs
• newSession
• setActiveTab
• getActiveTab
• listFrames
• uploadFile
• dropFile
• reload
• fork
• killFork
• listForks
• parallel
• tryUntil