Chrome extension

rly — remote browser for AI agents

Install this in the Chrome you want an AI agent to use. One click creates a private session link; paste it into your agent and it can drive this browser. No local install, no admin rights, and it uses the Chrome you already have.

Install from the Chrome Web Store

Listing pending review. The extension has been submitted to the Chrome Web Store and is awaiting Google's review. This page will link to the listing once it is published. Until then, use the manual install below. It is the same build.
  1. Open the Chrome Web Store listing for rly — remote browser for AI agents (link coming soon) and click Add to Chrome.
  2. Pin the rly icon from the puzzle-piece menu so you can see its status.
  3. Continue with Start a session.

Manual install (unpacked)

Use this while the store listing is pending, or where the store is blocked but developer mode is allowed.

  1. Download rly-extension.zip and extract it to a folder you will keep, for example Documents\rly-extension. Chrome loads the extension from that folder every time it starts.
  2. Open chrome://extensions (or edge://extensions in Microsoft Edge).
  3. Turn on Developer mode (top right).
  4. Click Load unpacked and select the extracted folder, the one that contains manifest.json.
  5. Pin the rly icon from the puzzle-piece menu.

To update, extract the new zip over the same folder and click the reload icon on the rly card in chrome://extensions. If your organisation blocks developer mode or extensions, use rly-host.exe instead.

Start a session

  1. Click the rly icon and click Create session. The relay defaults to https://rly.ems.works; to use your own, change it under Relay settings first.
  2. The popup shows a Session link such as https://rly.ems.works/mcp/rly_…. Click Copy, or Copy Claude Code command for the ready claude mcp add command.
  3. Paste it into your AI coding agent. For Claude Code:
    $ claude mcp add --transport http rly https://rly.ems.works/mcp/rly_…
  4. When you are done, click End session or close the browser.

The agent controls the whole browser: every tab in every normal window, including tabs you open later. chrome:// pages, other extensions' pages and the Chrome Web Store are excluded.

You stay in control

Why it asks for these permissions

PermissionUsed for
debuggerDriving the browser's tabs during a session, using the Chrome DevTools Protocol.
tabsOpening, closing and listing the agent's tabs.
tabGroupsKeeping agent tabs in a labelled “rly agent” group.
storageKeeping the session token in session storage (wiped when the browser closes) and the relay setting locally.
alarmsKeeping the relay connection alive and reconnecting.

The extension has no content scripts and requests no host permissions.