Tabrelay / Playwright and Puppeteer with a logged-in Chrome
Puppeteer and Playwright with your logged-in Chrome: an alternative
Puppeteer, Playwright and Selenium are excellent, but they drive a separate browser. If what you want is to automate sites where you are already logged in, there is a simpler path.
Chrome or Chromium, Python 3.10+, macOS and Linux. Source on GitHub.
What Puppeteer, Playwright and Selenium assume
These tools normally start a browser of their own, in a clean or dedicated profile. That is the right design for tests and for anything running on a server. For your own day-to-day accounts it means logging in again in that browser, dealing with 2FA and "new device" checks, and keeping a separate profile alive. Some sites also treat automation browsers with suspicion.
Attaching to your everyday Chrome
Puppeteer and Playwright can connect to a Chrome that was started with --remote-debugging-port. It works, but it has costs:
- A debugging port on your everyday browser gives every local program full control of every site, with no per-site limit.
- Recent Chrome versions (from Chrome 136) ignore the remote debugging switch for the default profile directory, so you have to start Chrome with a separate
--user-data-dirand log in again. Check Chrome's current documentation, as this has changed over time. - You have to start Chrome yourself, in a special way, every time.
The Tabrelay approach
Tabrelay runs inside the browser you already use. A Chrome extension carries out jobs with chrome.debugger, only on sites on an allow-list that you edit on the extension's own options page, and only for callers that hold a local secret token. The command line talks to a small server on 127.0.0.1 that refuses foreign Host headers and web page origins.
| Playwright / Puppeteer | Tabrelay | |
|---|---|---|
| Browser | starts its own, or attaches to a debug port | your normal Chrome profile, via an extension |
| Logins and 2FA | log in again in that browser | already there |
| Site limits | none by default | allow-list in the extension |
| Language | Node, Python, Java, .NET (Playwright); Node (Puppeteer) | CLI, plus a Python module |
| Locators, auto-waiting | yes | no; use wait_for and JavaScript |
| Browsers | Playwright: Chromium, Firefox, WebKit | Chrome and Chromium only |
| Headless, servers, CI | yes | not what it is for |
When Tabrelay is the wrong choice
Use Playwright or Puppeteer for tests, scraping at scale, headless runs, anything on a server, and anything that needs other browsers. Tabrelay does less on purpose: no auto-waiting locators, no browser management, Chrome and Chromium only, and a page can tell that a debugger is attached.
When Tabrelay fits
- Small jobs in accounts you are logged in to: posting, uploading, exporting, checking a page.
- Keeping a login alive with
keepalive.py, which visits a site in an inactive background tab on a schedule. - Letting a script or an AI agent use your session without handing it your password.
See also Tabrelay vs Playwright MCP, browser-use and Claude in Chrome. Get Tabrelay from the latest release.
Frequently asked questions
Can Playwright use my logged-in Chrome profile?
It can attach to a Chrome started with a remote debugging port, or launch a persistent profile of its own. Neither is quite your everyday Chrome profile, and recent Chrome versions restrict remote debugging on the default profile. Check the current Playwright and Chrome documentation.
Is Tabrelay a replacement for Playwright?
No. It covers a narrow case, scripting your own logged-in Chrome, and does less: no auto-waiting locators, no browser management, and Chrome or Chromium only.
Is a remote debugging port safe?
Any local program that can reach the port can control every site in that browser. Tabrelay avoids opening one: it uses an extension with an allow-list and a local token.
Does Tabrelay need Node?
No. It needs Chrome and Python 3.10 or newer, with only the standard library.
Chrome or Chromium, Python 3.10+, macOS and Linux. Source on GitHub.