Repolode
Submit
browseros-ai

browseros-ai/BrowserOS

An open-source Chromium-based browser built for AI agents, which runs locally and lets agents work in parallel tabs using your logged-in accounts.

Open-source alternative to:ChatGPT AtlasPerplexity CometDia

View on GitHubWebsite
Stars
14,006
Forks
1,504
Open issues
99
Last commit
18 hours ago

What is it?

BrowserOS neo is a secondary browser designed for AI agents rather than people. It imports logins, bookmarks and extensions from Chrome in one click, so agents can work inside accounts you already use, such as posting to social media, clearing an inbox, or pulling reports from internal tools.

Agents connect through MCP. The project detects Claude Code, Codex, Cursor, VS Code and several other tools on the machine and links them with one click. Each agent runs in its own tab, a live dashboard shows current sessions, and every session is saved to disk as a scrubbable video with an action timeline.

All sessions, screenshots and history stay under a local directory and run on 127.0.0.1. Anonymous usage events can be turned off in Settings. The repository also contains BrowserOS, a Chromium fork with a built-in AI agent for people who browse themselves.

Key features

  • 01One-click import of logins, bookmarks and extensions from Chrome
  • 02Connects to MCP-compatible agents such as Claude Code, Codex and Cursor
  • 03Runs multiple agents in parallel, each in its own tab
  • 04Live dashboard of active agent sessions
  • 05Sessions saved as replayable videos with an action timeline
  • 06Runs locally on macOS, Windows and Linux x64

Use cases

  • Letting an agent post to LinkedIn or Twitter/X from your real logged-in account
  • Automating inbox cleanup and unsubscribing from newsletters
  • Updating a CRM or filing expenses through web interfaces with an AI agent
  • Auditing what an agent did on a website by replaying its session

Quick start

  1. 1

    Install BrowserOS neo with Homebrew

    brew tap browseros-ai/tap && brew install --cask browseros-neo
  2. 2

    Or download the app for your platform from the release links

  3. 3

    Import your logins from Chrome in one click

  4. 4

    Connect your AI agent (Claude Code, Codex, Cursor, etc.) from the app

Things to know

  • Chromium-based build; the Linux builds are x64 only.
  • Agents act on your real logged-in accounts, so mistakes can affect your live accounts.
  • Full source build of the browser needs about 100GB of disk space.

Topics

#ai-agents#browser-automation#mcp#chromium#browser#llm#automation#agent#browseros#claude

FAQ

What is BrowserOS?
An open-source Chromium-based browser built for AI agents, which runs locally and lets agents work in parallel tabs using your logged-in accounts. BrowserOS neo is a secondary browser designed for AI agents rather than people. It imports logins, bookmarks and extensions from Chrome in one click, so agents can work inside accounts you already use, such as posting to social media, clearing an inbox, or pulling reports from internal tools.
Is BrowserOS free to use? What license does it have?
BrowserOS is open source under the AGPL-3.0 license. Check the license terms before using it in commercial products.
Is BrowserOS still maintained?
Actively maintained. The last commit to BrowserOS was on Oct 19, 2026.
What can I use BrowserOS for?
Letting an agent post to LinkedIn or Twitter/X from your real logged-in account Automating inbox cleanup and unsubscribing from newsletters Updating a CRM or filing expenses through web interfaces with an AI agent Auditing what an agent did on a website by replaying its session
How do I install BrowserOS?
Install BrowserOS neo with Homebrew: brew tap browseros-ai/tap && brew install --cask browseros-neo Or download the app for your platform from the release links Import your logins from Chrome in one click Connect your AI agent (Claude Code, Codex, Cursor, etc.) from the app
Where can I get BrowserOS?
The source code and installation instructions are on GitHub: https://github.com/browseros-ai/BrowserOS

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects