• 🤖 Use different chatbots in one app, currently supporting ChatGPT, new Bing Chat, Google Bard, Claude, and open-source models including LLama2, Vicuna, ChatGLM etc
• 💬 Chat with multiple chatbots at the same time, making it easy to compare their answers
• 🚀 Support ChatGPT API and GPT-4 Browsing
• 🔍 Shortcut to quickly activate the app anywhere in the browser
• 🎨 Markdown and code highlight support
• 📚 Prompt Library for custom prompts and community prompts
• 💾 Conversation history saved locally
• 📥 Export and Import all your data
• 🔗 Share conversation to markdown
• 🌙 Dark mode
• 🌐 Web access

• ChatGPT (via Webapp/API/Azure/Poe)
• Bing Chat
• Google Bard
• Claude 2 (via Webapp/API/Poe)
• LLaMA 2
• ChatGLM
• Pi by Inflection
• Vicuna
• WizardLM
• iFlytek Spark
• Tongyi Qianwen
• Baichuan
• ...

• Clone the source code
• corepack enable
• yarn install
• yarn build
• In Chrome/Edge go to the Extensions page (chrome://extensions or edge://extensions)
• Enable Developer Mode
• Drag the dist folder anywhere on the page to import it (do not delete the folder afterward)

ChatHub can expose a local OpenAI-compatible HTTP API so VS Code extensions like Cline, Roocode, and Kilocode can connect to your configured providers via the browser extension.

What you get:

• HTTP endpoints on http://127.0.0.1:4891
  • GET /v1/models
  • POST /v1/chat/completions (supports streaming via SSE)
  • POST /v1/embeddings
• Requests are routed through ChatHub, reusing your configured providers and sessions.

Setup

1. Build and load the extension

   • corepack enable
   • yarn install
   • yarn build
   • Load dist into Chrome/Edge as an unpacked extension

2. Install the Native Host (provides the local HTTP server)

   • cd native-host
   • chmod +x install.sh
   • ./install.sh EXTENSION_ID
     • Replace EXTENSION_ID with your loaded extension’s ID from chrome://extensions
     • This installs:
       • /usr/local/bin/chathub-native-host (the server)
       • Native messaging manifest under the Chrome native hosts directory

3. Enable the proxy in the extension settings

   • Open ChatHub, go to Settings
   • Ensure your provider keys/sessions are configured (e.g. OpenAI)
   • The native bridge starts automatically; the local server listens on 4891 by default

Quick tests

• curl -s http://127.0.0.1:4891/v1/models
• curl -s -H "Authorization: Bearer test" -H "Content-Type: application/json" -d '{"model":"gpt-3.5-turbo","messages":[{"role":"user","content":"hello"}]}' http://127.0.0.1:4891/v1/chat/completions
• Streaming:
  • curl -N -s -H "Authorization: Bearer test" -H "Content-Type: application/json" -d '{"model":"gpt-3.5-turbo","stream":true,"messages":[{"role":"user","content":"hi"}]}' http://127.0.0.1:4891/v1/chat/completions

VS Code client configuration

• Base URL: http://127.0.0.1:4891
• API Key: any placeholder (e.g., chathub-local). If you have an actual OpenAI key set in ChatHub settings, it will be used automatically.
• Cline/Roocode/Kilocode should work without additional changes.

Notes

• The extension requires “Native Messaging” permission.
• The local server currently supports OpenAI-compatible endpoints. More providers/routes will appear as you configure them inside ChatHub.
• For dev builds, extension ID changes; re-run the installer if the ID changes.