Aipex

Your browser already works!

AIPex — An open-source browser automation agent that lives in your existing browser.

• Zero Migration: No new browser to install. No new workflow to learn.
• Open Source: MIT licensed. Fully transparent, auditable, and extensible.
• Privacy First: Your data never leaves your machine. Bring Your Own Key (BYOK).

Why We Built This

Every browser automation tool asks you to:

• Install a separate browser (Dia/Comet)
• Pay monthly subscriptions (ChatGPT Atlas)
• Give up your browsing data

We asked: why can't automation just run in the browser you already use?

AIPex is the answer. Install the extension, bring your own API key, and automate anything — right where you already work.

Quick Start

1. Install — Chrome Web Store or Edge Add-ons
2. Open — Press AIPex icon
3. Automate — Type or speak what you want in natural language

Use with AI Coding Agents (MCP)

AIPex now supports the Model Context Protocol (MCP), giving AI coding agents like Cursor, Claude Code, and VS Code Copilot direct control over your browser.

AI Agent ──stdio──▶ aipex-mcp-bridge ──WebSocket──▶ AIPex Extension ──▶ Browser

Step 1: Configure your agent

Cursor (.cursor/mcp.json) · Claude Desktop (claude_desktop_config.json) · Windsurf (mcp_config.json):

{
  "mcpServers": {
    "aipex-browser": {
      "command": "npx",
      "args": ["-y", "aipex-mcp-bridge"]
    }
  }
}

Claude Code:

claude mcp add aipex-browser -- npx -y aipex-mcp-bridge

VS Code Copilot (.vscode/mcp.json):

{
  "servers": {
    "aipex-browser": {
      "command": "npx",
      "args": ["-y", "aipex-mcp-bridge"]
    }
  }
}

Step 2: Connect the extension

1. Open Chrome → AIPex icon → Options
2. Set WebSocket URL to ws://localhost:9223
3. Click Connect

Your agent now has 30+ browser automation tools available via MCP. See mcp-bridge/README.md for advanced options.

Skill

AIPex ships an aipex-browser skill — a ready-to-use skill package for agents that support the skill protocol (such as Claude Code and OpenClaw-compatible runtimes).

The skill bundles tool usage strategy, complete parameter schemas for all 30+ browser tools, and common automation patterns — so an agent can control the browser effectively without discovering tools from scratch.

See skill/SKILL.md for the full skill definition.

Demos

"I have 100 tabs open. Help."

https://github.com/user-attachments/assets/4a4f2a64-691c-4783-965e-043b329a8035

"Research this topic without leaving my browser"

https://github.com/user-attachments/assets/71ec4efd-d80e-4e8f-8e39-88baee3ec38e

"Write a tweet for me"

https://github.com/user-attachments/assets/81f6b482-84d0-4fd9-924b-dca634b208ec

"Help me pass this exam"

https://github.com/user-attachments/assets/ba454715-c759-41df-bf87-e835f76be365

Roadmap?

• Page Understanding

  • Accessbility Tree

  • Optimised Dom

  • Vision

• Context Engineering

  • Search-based Retrival

  • Drop unused snapshot

  • id-based operation

• Integration

  • Cursor

  • Claude Code

• Skills

  • File system

  • Script Execution

• Evaluation - Online-Mind2Web

Why debugger is necessary for browser automation?

Contributing

We love contributions! See DEVELOPMENT.md for setup instructions.

Contributors

Star History