a CLI tool that gives any AI coding agent the ability to control Chrome. open pages, click buttons, type text, take screenshots, extract content — all from the terminal.
AI agents are powerful, but they can't see or interact with the web by default. this tool changes that. once installed, your agent can browse, automate, and scrape anything you can do in a browser.
compiled CLI — ready to use, no building required
chrome launcher — script to start Chrome with remote debugging (macOS + Linux)
image uploader — helper for uploading files to any web page
setup guide — step-by-step instructions + agent integration examples
node dist/browser.js list # list open tabs node dist/browser.js open <url> --new-tab # open a URL node dist/browser.js content <tab> # get page text node dist/browser.js click <tab> "Submit" # click a button node dist/browser.js type <tab> "hello" # type into an input node dist/browser.js screenshot <tab> -o f # take a screenshot node dist/browser.js elements <tab> # list interactive elements node dist/browser.js html <tab> # get raw HTML node dist/browser.js search <query> # search across tabs node dist/browser.js close all # clean up tabs
1. unzip the bundle
unzip browser-agent.zip2. install dependencies
cd browser-agent-bundle && npm install3. start Chrome with debugging
chmod +x scripts/start-chrome.sh && ./scripts/start-chrome.sh4. verify it works
node dist/browser.js listNode.js 18+
Google Chrome (any recent version)
macOS or Linux (Windows via WSL)
> social media automation — post, comment, engage on any platform
> web scraping — extract data from any page Claude can see
> form filling — automate signups, checkouts, data entry
> testing — let your agent QA your web app by actually using it
> monitoring — check dashboards, track prices, watch for changes
instant download after payment. 3 download attempts. no subscription.