Levr Browser

Let your AI agents test your web app in real browser tabs, with every screenshot and result recorded in Levr.

It's not "done" until it's tested. Levr Browser lets the agents you use with Levr (Claude Code, Codex, Cursor and others) drive real browser tabs to test your web application: open pages, click, type, fill forms, take screenshots, and read console and network logs. Screenshots and results are recorded in your Levr test runs and issues, so you see what the agent actually tested, not just what it says it did. How it works - Sign in to your Levr account from the side panel. - Agents reach your browser only through your Levr workspace. They connect to the levr-browser MCP server on the Levr API, and Levr routes each request to your signed-in browser. Nothing listens on your computer. - Agents work only in the tabs of the "Levr" tab group: tabs they open, or tabs you choose to share. Your other tabs are not touched. - Stop all, in the side panel, stops every agent at once. Agents cannot resume on their own. What agents can do - Navigate, open and close tabs and windows in the Levr tab group. - Click, type, press keys, hover, scroll, drag and fill forms. - Take screenshots and record short screencasts, saved to your Levr workspace. - Read console messages, network requests and page content. - Run JavaScript in a page of the Levr tab group, as the DevTools console does, to inspect or set up the page under test. - Emulate a device or screen size. What agents cannot do - Act on Levr's own sign-in, consent and workspace pages, or on the browser's internal pages. - Read files from your computer, except a folder you choose in Settings for file uploads. - Use the extension without your Levr sign-in. Levr Browser is part of Levr (https://levr.one), a platform for AI-assisted testing and making sure your LLM implements what you asked for.