• 100% Local - All processing happens on your machine. No data sent to external servers (except optional web search)
• Multi-Model Support - Load and switch between any models from LM Studio
• Real-time Streaming - See responses generate word-by-word
• Smart Web Search - Auto-detects when to search the web for current information
• Source Citations - Every web search result includes clickable source links
• Conversation Memory - Maintains context throughout your chat session

• Multi-Model Responses - Query multiple AI models simultaneously
• Color-Coded Chat - Each model gets a unique color for easy identification
• Group Chat Interface - Models respond together like a conversation
• Flexible Selection - Choose which models participate in the swarm
• Consensus Building - Compare different perspectives on the same question

• Modern React UI - Clean, responsive design with Tailwind CSS
• Dark/Light Themes - (Customizable via Tailwind)
• Model Selector - Easy dropdown to switch between loaded models
• Connection Status - Real-time indicator for LM Studio connection
• Mobile Responsive - Works great on any device

Ask questions and get intelligent responses powered by your local LLM:

You: What is quantum computing?
Bot: Quantum computing is a type of computing that uses quantum-mechanical
     phenomena...

Get multiple perspectives at once:

You: What's the best programming language for beginners?

🔵 Qwen-3: I'd recommend Python because of its readability...

🟣 Qwen-2.5: JavaScript is great for beginners since you can see results immediately...

🟢 Llama-3: Consider starting with Python for its extensive libraries...

Automatic web search for current information:

You: What's the weather in Tokyo today?
🔍 Searching the web...

Bot: According to current weather data [1], Tokyo is experiencing...

Sources:
[1] Weather.com - Tokyo Weather Forecast
[2] Japan Meteorological Agency - Current Conditions

1. Node.js (v16 or higher)

   node --version  # Should be 16.0.0 or higher

2. LM Studio - Download here

   • Install LM Studio
   • Download at least one model (recommended: Phi-3, Qwen, or Llama)

3. Brave Search API Key (Optional, for web search)

   • Free tier: brave.com/search/api
   • 2,000 queries/month for free

1. Clone the repository

   git clone https://github.com/yourusername/local-llm-chatbot.git
   cd local-llm-chatbot

2. Install dependencies

   npm install

3. Configure environment

   cp .env.example .env

   Edit .env and add your Brave API key (optional):

   BRAVE_API_KEY=your_brave_api_key_here

4. Set up LM Studio

   • Open LM Studio
   • Go to "Discover" and download a model
   • Load the model in the "Chat" tab
   • Go to "Local Server" → Click "Start Server"
   • Verify it's running on port 1234

5. Test the connection

   npm run test:lm

   You should see:

   ✅ LM Studio is working correctly!
   

6. Start the application

   npm run dev

7. Open in browser

   http://localhost:5173
   

1. Select a model from the dropdown
2. Type your question
3. Press Send or hit Enter
4. Watch the response stream in real-time

The chatbot automatically searches the web when you ask about:

• Current events, news, weather
• Latest versions, prices, stock data
• Real-time information
• Recent facts

Example queries:

• "What's the latest iPhone model?"
• "Current price of Bitcoin"
• "Weather in Paris today"

Activate Swarm Mode:

1. Click the "🐝 Swarm ON" button in the header
2. Select which models you want to participate
3. Ask your question
4. All selected models will respond with unique colors

Use Cases for Swarm Mode:

• Compare perspectives - See how different models approach the same problem
• Fact-checking - Cross-reference information across models
• Creative brainstorming - Get diverse ideas from multiple AI personalities
• Code review - Multiple models can spot different issues
• Learning - See how different models explain concepts

1. Add font files to public/fonts/
2. Update src/index.css:

   @font-face {
     font-family: 'YourFont';
     src: url('/fonts/YourFont.woff2') format('woff2');
   }
   body {
     font-family: 'YourFont', sans-serif;
   }

Simply load different models in LM Studio and they'll appear in the dropdown automatically!

Edit server.js line 109:

temperature: 0.3, // Lower = more focused, Higher = more creative

localLLM/
├── src/
│   ├── App.jsx              # Main React component with Swarm Mode
│   ├── main.jsx             # React entry point
│   └── index.css            # Tailwind styles
├── public/                  # Static assets
├── server.js                # Express backend with multi-model support
├── test-lmstudio.js         # LM Studio connection tester
├── .env.example             # Environment variables template
├── vite.config.js           # Vite configuration
├── tailwind.config.js       # Tailwind CSS configuration
└── package.json             # Dependencies and scripts

Send a message and receive streaming response

Request:

{
  "message": "Your question",
  "conversationId": "unique-id",
  "model": "qwen/qwen3-vl-4b",
  "swarmMode": false,
  "models": ["model1", "model2"]
}

Response: Server-Sent Events (SSE)

data: {"type":"status","status":"Analyzing query..."}
data: {"type":"content","content":"Response text...","modelId":"model-name"}
data: {"type":"sources","sources":[...]}
data: {"type":"done"}

Get list of available models from LM Studio

Response:

{
  "models": [
    {"id": "qwen/qwen3-vl-4b", "name": "qwen3-vl-4b"},
    {"id": "qwen2.5-3b-instruct", "name": "qwen2.5-3b-instruct"}
  ]
}

Check server and LM Studio connection status

Clear conversation history

Contributions are welcome! Please feel free to submit a Pull Request.

1. Fork the repository
2. Create your feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add some AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

• Ensure LM Studio is running
• Check that Local Server is started (port 1234)
• Load a model in LM Studio
• Run npm run test:lm to diagnose

• Make sure models are loaded in LM Studio (not just downloaded)
• Restart the dev server
• Check browser console for errors

• Verify BRAVE_API_KEY in .env
• Check free tier limits (2000 queries/month)
• Web search is optional - chatbot works without it

• Try smaller models (Phi-3, Qwen-2.5-3B)
• Close other applications
• Check CPU/GPU usage
• Reduce context length in LM Studio settings

npm run dev          # Start both backend and frontend
npm run dev:server   # Start backend only
npm run dev:client   # Start frontend only
npm run test:lm      # Test LM Studio connection
npm run build        # Build for production

• 100% Local Processing - Your conversations never leave your machine
• No External APIs - Except optional Brave Search (can be disabled)
• No Data Collection - We don't collect, store, or transmit your data
• Open Source - Full transparency, inspect the code yourself

This project is licensed under the MIT License - see the LICENSE file for details.

• LM Studio - For making local LLMs accessible
• Brave Search - For the free search API
• React - UI framework
• Tailwind CSS - Styling
• Vite - Build tool
• Express - Backend framework

If you find this project useful, please consider giving it a star!