YC S24 backed Chromium fork with native AI support. 314 HN points with 206 comments is strong organic engagement. Comments flagged prompt injection risk as a known concern.
BrowserOS / Nxtscape
activeChromium fork with native AI agent support. YC S24 backed. Privacy-first agentic browser with 12 built-in AI skills. Supports 11+ LLM providers including Ollama for local models. Only credible open-source alternative to ChatGPT Atlas, Perplexity Comet, and Dia.
Where it wins
Chromium fork with native AI agent integration — not a plugin or overlay
YC-backed with weekly releases: 47 releases in ~10 months, v0.43.0 on 2026-03-12
31-tool MCP server integration (v0.42.0); Skills, Memory, SOUL.md in v0.43.0
Supports 11+ LLM providers including local models (Ollama, LMStudio) — browsing history stays on device
314-point HN (Jun 2025) + 88-point thread (Jan 2026 as BrowserOS)
Only credible open-source alternative to Atlas/Comet/Dia
Where to be skeptical
~10 contributors despite YC backing — thin team for a browser product
AGPL-3.0 license may block enterprise adoption
Lane 4 (consumer agentic browser) — different market than developer automation tools
HN criticism flagged prompt injection risk
Editorial verdict
Lane 4 leader. Just crossed 10K stars. 314 HN pts with 206 comments is strong organic engagement. YC S24 backing. Privacy-first Chromium fork with built-in AI. Different market than Lanes 1-3 but clear leader in consumer agentic browsers.
Related
Playwright MCP
93Microsoft's official MCP server for Playwright. Uses accessibility snapshots instead of screenshots for structured browser control. Auto-configured in GitHub Copilot's Coding Agent.
Chrome DevTools MCP
92Google Chrome team's official MCP server for Chrome DevTools. Gives coding agents deep debugging, performance profiling, and Core Web Vitals analysis through 26 tools across 6 categories.
Stagehand
90AI-native browser automation SDK by Browserbase with natural language selectors and act/extract/observe primitives.
Skyvern
90Vision-LLM browser automation for enterprise workflows. Combines computer vision with LLM reasoning to handle websites never seen before. YC S23 backed with CAPTCHA solving, 2FA, and proxy networks.
Public evidence
YC S24 backing provides institutional credibility. Available on macOS, Windows, Linux. 12 built-in AI skills (deep research, price comparison, form filling).
Just crossed the 10K star threshold. Actively maintained — pushed Mar 19. Privacy-first, supports local models (Ollama). 47 releases in ~10 months.
v0.43.0 (2026-03-12) adds Skills, Memory, and SOUL.md — agent-native primitives uncommon in Chromium forks. v0.42.0 added MCP server integration with 31 tools. Weekly release cadence confirms active development.
Raw GitHub source
GitHub README peek
Constrained peek so you can sanity-check the source material without leaving the site.
<br></br> <br></br>
</div> <table> <tr> <td width="50%" align="center" valign="top"> <img src="https://raw.githubusercontent.com/nxtscape/nxtscape/main/packages/browseros/resources/browserclaw/icons/product_logo_192.png" alt="BrowserClaw logo" width="96" height="96" />BrowserClaw
The browser for AI agents
Claude Code, Codex, Cursor, or any MCP client drives it using the accounts you're already signed into, while you watch live and replay every step.
Website · Docs
</td> <td width="50%" align="center" valign="top"> <img src="https://raw.githubusercontent.com/nxtscape/nxtscape/main/packages/browseros/resources/browseros/icons/product_logo_192.png" alt="BrowserOS logo" width="96" height="96" />BrowserOS
The AI browser for humans
An open-source Chromium fork with a built-in AI agent, the privacy-first alternative to ChatGPT Atlas, Perplexity Comet, and Dia.
Website · Docs
</td> </tr> </table> <div align="center">Two browsers, one codebase. Free · Open source (AGPL-3.0) · Local-only · Bring your own AI keys
</div>BrowserClaw
What is BrowserClaw? BrowserClaw is a free, open-source browser your AI agents drive using your logged-in accounts. You install it like any browser, sign into the sites you use, and give tasks to Claude Code, Codex, Cursor, or any MCP-compatible AI. Your agents work in their own tabs while you watch every step live and replay any session like a video.
Your AI is smart, but it can't press the buttons. Ask it to book a flight, download an invoice, or reply to an email, and it stops at the login screen. BrowserClaw fixes that.
Get started
- Install BrowserClaw and sign in to the sites you use every day. It works like any browser, and every account you sign into becomes something your AI can use.
- Connect your AI in one click. Claude Code, Codex, Cursor, VS Code, Zed, OpenCode, and Antigravity install with a single click; anything else that speaks MCP connects with a URL.
- Give it a real task. In your AI chat: "Find a good time next week for a 30-minute team meeting and send the invite." Then watch it live from your new tab and replay the whole run like a video.
Key features
<table> <tr> <td width="40%" valign="middle"> <h4>Live dashboard</h4> Your new tab shows every agent working right now: which site it's on, what it's doing, how far along. <a href="https://docs.browseros.com/browserclaw/cockpit">Docs</a> </td> <td width="60%"> <img src="https://raw.githubusercontent.com/nxtscape/nxtscape/main/docs/images/browserclaw--dashboard-populated.png" alt="BrowserClaw dashboard showing agent sessions and recent activity" width="100%" /> </td> </tr> <tr> <td width="40%" valign="middle"> <h4>One-click MCP connect</h4> One endpoint, every harness. Seven AI tools set up with a single click. <a href="https://docs.browseros.com/browserclaw/mcp">Docs</a> </td> <td width="60%"> <img src="https://raw.githubusercontent.com/nxtscape/nxtscape/main/docs/images/browserclaw--mcp-install-board.png" alt="BrowserClaw MCP connect board with one-click install for supported AI tools" width="100%" /> </td> </tr> <tr> <td width="40%" valign="middle"> <h4>Replay every session</h4> Every session is saved as a scrubbable video on your disk with a step-by-step action timeline. Rewind and see exactly what happened. <a href="https://docs.browseros.com/browserclaw/audit-and-replay">Docs</a> </td> <td width="60%"> <img src="https://raw.githubusercontent.com/nxtscape/nxtscape/main/docs/images/browserclaw--replay-scrubber.png" alt="BrowserClaw replay view with video scrubber and action timeline" width="100%" /> </td> </tr> </table>- Your logins. Agents automate your real work using the sessions you already have, not a blank sandbox. How it works
- Local-only. Sessions, screenshots, and history live under
~/.browserclaw/and never leave your machine. Privacy
Why BrowserClaw over the alternatives?
- Not a headless driver. Playwright and browser-use spin up a fresh Chrome subprocess with no logins. Great for CI, useless for "book my flight" or "read my inbox." BrowserClaw is the browser your logins already live in.
- Not a cloud browser. Browserbase and Browser Use Cloud run your agent's session on someone else's infrastructure. Your prompts and session tokens pass through their servers. BrowserClaw runs on your machine, on
127.0.0.1.
BrowserOS
What is BrowserOS? BrowserOS is a free, open-source Chromium fork with an AI agent built into every new tab. Ask it to summarise a page, click through a flow, extract data, or run a scheduled task, and it uses 20+ built-in tools plus 40+ app integrations to get the work done. Bring your own AI keys or run everything locally with Ollama.
Every AI browser today asks you to sign into their cloud and hand over your data. BrowserOS is the one that doesn't. Same daily browser you already use, with a helpful agent one keystroke away.
Get started
- Download and install BrowserOS: macOS · Windows · Linux (AppImage) · Linux (Debian).
- Import from Chrome in one click. Bookmarks, passwords, extensions all carry over.
- Connect your AI provider. Claude, OpenAI, Gemini, ChatGPT Pro via OAuth, or local models via Ollama or LM Studio.