browsentic

scrapingimshaikot30 stars

An agentic browser extension: an AI side panel in your own logged-in Chrome, Edge, Arc, Brave or even Firefox run by the agent you already use (Claude Code, Codex, Antigravity). Free, open source, no API key. Also an MCP server.

Install

npx -y browsentic

README.md

https://github.com/user-attachments/assets/e9f146d3-cb14-416e-b8d7-68c4e81cbd0d

Browsentic – a completely integrated agentic browser extension

No account · No auth · No API key · No cloud · No telemetry · Local first

Your real, logged-in browser, driven from an AI side panel by the agent CLI you already have.
Also an optional MCP server, so any MCP client can drive the same browser.

Works in

Runs on

Quick Start

Any platform — with Node.js 20 or newer (Windows is experimental):

npx browsentic setup

macOS — one line installs Browsentic.app, which brings Node and everything else:

curl -fsSL https://browsentic.com/install.sh | sh

Windows — one line in PowerShell installs the Windows app, which brings Node and everything else (experimental):

irm https://browsentic.com/install.ps1 | iex

Full install and setup guide, Firefox included: browsentic.com/docs/guide/install

Key Capabilities

🧭 Side panel, human in the loop	🤖 Every major agent CLI	🎯 50+ deterministic tools
🌐 All Chromium + Firefox	🗺️ Full browser navigation	🧩 Solves captchas on the way
📄 Reads files, acts on them	↕️ Uploads & downloads	🧠 Creates its own skills
⚡ Writes & runs live tools	🎙️ Voice, text or show it once	👁️ Point at what you mean
⏱️ Monitors & schedules	🛡️ Guardrails & approvals	🔌 Optional MCP server

How It Works

How an instruction becomes a click

The same flow, as text

You ──speak or type──> Extension ──local WebSocket──> Daemon ──spawns──> your agent CLI
                            ▲                              (claude │ codex │ agy │ vibe │ grok)
                            └──────────────── page actions ─────────────────────┘

Any MCP client ──stdio──> browsentic mcp ──> the same daemon ──> the same browser

The extension dials out to the daemon, because a Manifest V3 service worker cannot listen for connections. One daemon owns the browser link, so several MCP clients can share one browser. Everything binds to 127.0.0.1.

Resources

📚 Documentation
🚀 Install and Pair
✨ Features
🧰 All 52 Page Tools
🔌 Optional: Drive It From Claude Code, Cursor or Zed Over MCP
🛡️ Approvals and Guardrails
🏗️ Architecture
🩺 Troubleshooting

Privacy and Security

Nothing connects until you pair, both ends prove themselves, consequential actions ask first, and credentials on a page are sealed before the agent sees them. The full model is in SECURITY.md, and what it does not cover is in Limits.

Contributing

Bugs and ideas are welcome — start at CONTRIBUTING.md.

License

Browsentic is MIT licensed.

Source Available: Always visible source code
Local First: No cloud component, no telemetry, no account
Extensible: Add your own page capabilities, skills and agent runners

What does Browsentic mean?

Short answer: "Browse" + "agentic".

Long answer: Most browser automation asks you to hand the work to a different browser — a headless one, in a container, logged in to nothing. Browsentic is the other way round: the agentic part happens in the browser you are already looking at, with the sessions you are already signed in to. The name is the thesis — browsing, made agentic, where you already browse.