[!NOTE] CoBridge also supports the companion browser extension Voyager, which is specifically designed for Gemini. If you only use Gemini, Voyager is a lightweight alternative. For broader platform support (ChatGPT, Claude, DeepSeek, Doubao), use the CoBridge browser extension.

✨ Features

📝 Smart Context Extraction

Extract AI conversation context from web pages, including text, tables, and images

🎯 System Prompt Completion

Automatically fill in system prompts for workspace Agents, quickly bootstrapping projects

🌿 Wide Ecosystem Support

Supports ChatGPT, Claude, DeepSeek, Doubao and other AI platforms

Supports GitHub Copilot, Cursor, Claude Code and other IDE Agents

📥 Installation

Store Install

Step 1: Install the Browser Extension

Install from the Chrome Web Store (coming soon)

Step 2: Install the VS Code Extension

Open the Open VSX Marketplace, search for CoBridge, and click install.

Manual Install

Install the Browser Extension

1. Download or clone this repository
2. Run pnpm install && pnpm --filter chrome-extension build in the project root
3. Open chrome://extensions/, enable Developer Mode
4. Click Load unpacked and select packages/chrome-extension/dist

🏗️ Architecture

┌─────────────────────┐          ┌─────────────────────┐
│   Browser Extension  │  HTTP    │   VS Code Extension  │
│   (Chrome / Edge)    │ ──────► │   (Agent Receiver)   │
│                      │ :3030   │                      │
│  Captures AI chat    │         │  Saves to workspace  │
│  from web pages      │         │  .cobridge/CONTEXT   │
└─────────────────────┘          └─────────────────────┘

• Browser Extension: Injects into AI chat pages, extracts conversation content (text, tables, images), and sends it to localhost
• VS Code Extension: Runs a local HTTP server, receives the data, and writes it as Markdown context files that your IDE Agent can read

🚀 Quick Start

Step 1: Select Your Agent

Click the status bar icon to open the management menu:

Select the Agent you're currently using:

Once selected, the status bar displays the corresponding Agent icon:

Step 2: Start the Service

Click Start Service from the menu. The local server will run on port 3030 until you stop it.

Step 3: Sync Your Conversations

1. Open any supported AI chat page in your browser
2. Click the CoBridge browser extension icon
3. Hit Sync Context to Agent
4. The conversation content lands in .cobridge/AI_CONTEXT.md in your workspace

From now on, your Agent will never look at you blankly and ask, "What did you say before?"

⚙️ Configuration

Change the Port

If port 3030 is occupied:

1. Open VS Code Settings (Ctrl + , / Cmd + ,)
2. Search for AIContextSync.port
3. Change the port number (e.g., 3031)
4. Restart the service from the status bar menu

Note: VS Code workspace settings override user settings. Modify the port in your Workspace Settings if needed.

The browser extension's popup also allows you to set the port to match.

🌐 Supported AI Platforms

🤖 Supported Agents

🎯 Principles

• Zero Pollution: CoBridge automatically adds the sync file to .gitignore, ensuring it never pollutes your Git repository.
• Friendly Format: Full Markdown output — tables, images, and text all preserved in a format your IDE Agent can read naturally.
• Auto Configuration: Automatically generates rule files adapted to various Agents, allowing them to seamlessly read context without bloat.
• Local First: All data stays on your machine. No external servers, no telemetry.

📁 Project Structure

This is a pnpm workspace monorepo containing two packages:

CoBridge/
├── packages/
│   ├── chrome-extension/     # Browser extension (Chrome / Edge)
│   └── vscode-extension/     # VS Code extension (Agent receiver)
├── pnpm-workspace.yaml
└── package.json

🛠️ Development

Prerequisites

• Node.js >= 20
• pnpm >= 8

Setup

git clone https://github.com/Winddfall/CoBridge.git
cd CoBridge
pnpm install

Build the Browser Extension

pnpm --filter chrome-extension build

Build the VS Code Extension

pnpm --filter vscode-extension compile

🌟 One Sentence Summary

Think in the browser. Land in the editor. Your Agent, forever equipped with cross-tab long-term memory.

If this project helped you, please give us a Star ⭐ on GitHub.

💡 Issues

If you have new requirements, welcome to raise an issue on GitHub.

🤝 Contributing

If you have good suggestions or find a bug, Pull Requests are welcome!

🥤 Sponsor This Project

If this project solved your AI collaboration pain points, feel free to buy me a cola! 🥤

Your support will be directly used to maintain subsequent iterations of the project ❤️.

Sponsor via WeChat / Alipay / Afdian:

WeChat Pay

Alipay

Afdian / 爱发电 Afdian

📄 License

This project is licensed under the MIT License.

Star History