browseros-ai/BrowserOS

🌐 The open-source Agentic browser; alternative to ChatGPT Atlas, Perplexity Comet, Dia.

What it solves

BrowserOS neo is a secondary browser designed specifically for AI agents to operate within. It solves the problem of agents lacking access to the user's authenticated sessions (logins), avoiding the limitations of headless drivers (which start blank) and cloud browsers (which are often blocked by sites like LinkedIn or X/Twitter due to datacenter IPs).

How it works

It is a Chromium fork that allows users to import their Chrome logins, bookmarks, and extensions in one click. It provides an MCP (Model Context Protocol) endpoint that AI agents (such as Claude Code, Codex, or Cursor) can connect to. Agents run in parallel in their own tabs, allowing the user to watch their progress live or replay sessions as scrubbable videos with a step-by-step action timeline.

Who it’s for

Users who employ AI agents to automate web tasks that require authenticated accounts, such as managing social media, clearing inboxes, or updating CRMs, and who want to keep these agentic activities separate from their primary browsing session.

Highlights

  • Authenticated Automation: Uses real logged-in accounts rather than a sandbox, enabling agents to perform tasks on the user's behalf.
  • Parallel Execution: Multiple agents can run simultaneously in separate tabs.
  • Local-first Privacy: All sessions, screenshots, and history are stored locally on the user's machine.
  • Broad Agent Compatibility: Works with any AI tool that speaks MCP, including Claude Code, Codex, Cursor, and VS Code.
  • Session Replay: Saves every session as a video with a timeline for easy auditing.

Related

  • Project
    platonai/Browser4

    An AI-native browser engine for autonomous agents and large-scale web automation, featuring zero-token ML extraction and high-performance swarm crawling.

  • Project
    LvcidPsyche/auto-browser

    An MCP-native browser control plane that gives AI agents a real browser with human-in-the-loop takeover, reusable auth profiles, and built-in safety rails.

  • Project
    anomalyco/browser-control

    A tool that lets AI agents control a user's existing Chromium-family browser using Playwright, allowing them to leverage active sessions and extensions.

  • Project
    risa-labs-inc/BossConsole

    BossConsole is an open‑source, JVM‑native desktop console that lets any LLM agent work with a real browser, terminal, editor, git, Docker/Kubernetes, and more. It exposes ~100 tools via a Model Context Protocol, offers fine‑grained RBAC and per‑tool kill‑switches, and supports hot‑reloading plugins so agents can evolve their own capabilities.