MultiBrain

MultiBrain is a powerful web application that queries multiple AI models simultaneously, providing real-time streaming responses and intelligent summaries. Built with Svelte 5 and FastAPI, it offers a modern, responsive interface for comparing AI responses side-by-side.

🚀 Features

• Multi-Model Querying: Query multiple AI providers simultaneously (OpenAI, Anthropic, Groq, Ollama, etc.)
• Real-Time Streaming: Watch responses stream in real-time from all configured models
• Intelligent Summaries: Automatic synthesis of all responses into a unified, fact-checked summary
• Bring Your Own Keys: Secure, client-side API key management with local encryption
• Modern UI: Beautiful, responsive interface built with Svelte and Tailwind CSS
• Keyboard Shortcuts: Efficient navigation with customizable keyboard shortcuts
• No Limits: Configure unlimited AI models with no artificial restrictions

📋 Requirements

• Python 3.8+
• Node.js 16+
• npm or yarn
• Svelte 5.0+ (included in dependencies)

🛠️ Installation

Backend Setup

1. Clone the repository:

git clone https://spacecruft.org/deepcrayon/multibrain
cd multibrain

2. Create a Python virtual environment:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install the package:

pip install -U pip setuptools wheel
pip install -e .

Frontend Setup

1. Navigate to the frontend directory:

cd frontend

2. Install dependencies:

npm install

🚦 Running the Application

Development Mode

1. Start the backend API server:

multibrain-api

The API server will run on http://localhost:8000

2. In a new terminal, start the frontend development server:

cd frontend
npm run dev

The frontend will run on http://localhost:5173

Production Mode

1. Build the frontend:

cd frontend
npm run build

2. Start the backend with static file serving:

multibrain-api --serve-static

Visit http://localhost:8000 to use the application.

🔧 Configuration

Adding LLM Providers

1. Click the "Configure LLMs" button in the header
2. Select a provider preset or choose "Custom"
3. Enter your API credentials:
   • Display Name: A friendly name for the LLM
   • API URL: The OpenAI-compatible endpoint
   • Model: The specific model to use
   • API Key: Your provider's API key

Supported Providers

• OpenAI: GPT-4, GPT-3.5-turbo
• Anthropic: Claude 3 Opus, Sonnet, Haiku
• Groq: Mixtral, Llama 2
• Ollama: Local models (Llama 2, Mistral, CodeLlama)
• Custom: Any OpenAI-compatible endpoint

Setting a Summary LLM

The summary LLM analyzes all responses and provides a unified answer:

1. Configure at least one LLM
2. In the configuration modal, select your preferred model from the "Summary LLM" dropdown
3. The summary will automatically generate after all responses complete

⌨️ Keyboard Shortcuts

• Ctrl+K: Open LLM configuration
• Ctrl+/: Focus query input
• Ctrl+Enter: Submit query
• Shift+?: Show keyboard shortcuts
• Escape: Close modals

🏗️ Architecture

Frontend (Svelte 5 + Vite)

The frontend has been fully migrated to Svelte 5, leveraging the latest features:

• Runes API: Using $state, $derived, $effect for reactive state management
• Props with $props(): Modern component prop handling
• Bindable Props: Two-way binding with $bindable
• Snippets: Reusable template fragments for better composition

frontend/
├── src/
│   ├── components/      # Svelte 5 components with runes
│   ├── stores/          # State management (.svelte.js files)
│   ├── lib/             # Utilities and helpers
│   └── App.svelte       # Main application

Key Svelte 5 Features Used:

1. State Management: All stores use .svelte.js extension with runes
2. Component Props: Migrated from export let to $props()
3. Reactive Effects: Using $effect for side effects instead of $:
4. Performance: Leveraging Svelte 5's improved reactivity system

Backend (FastAPI)

src/multibrain/
├── api/
│   ├── routes/          # API endpoints
│   │   ├── streaming.py # SSE streaming logic
│   │   └── router.py    # Route definitions
│   └── main.py          # FastAPI application

🔒 Security

• Local Storage: API keys are encrypted using Web Crypto API
• No Backend Storage: Keys are never sent to or stored on the backend
• Secure Proxy: Backend proxies requests to avoid CORS issues
• HTTPS Ready: Full support for secure deployments

🧪 Development

Svelte 5 Migration

This project has been fully migrated to Svelte 5. Key changes include:

• All components use the new runes API ($state, $derived, $effect)
• Stores are now .svelte.js files with reactive primitives
• Props use the $props() rune for better type safety
• Two-way binding with $bindable for form inputs
• Improved performance with Svelte 5's optimized compiler

Running Tests

# Backend tests
pytest

# Frontend tests
cd frontend
npm test

Component Development

When creating new components, follow the Svelte 5 patterns:

// Use $props() for component props
let { value = $bindable(''), onsubmit } = $props();

// Use $state for reactive state
let count = $state(0);

// Use $derived for computed values
let doubled = $derived(count * 2);

// Use $effect for side effects
$effect(() => {
  console.log('Count changed:', count);
});

Building for Production

# Build Python package
python -m build

# Build frontend
cd frontend
npm run build

📝 API Documentation

Once the backend is running, visit:

• Interactive API docs: http://localhost:8000/docs
• ReDoc: http://localhost:8000/redoc

Key Endpoints

• POST /api/query/stream: Stream responses from multiple LLMs
• POST /api/llm/validate: Test LLM configuration
• POST /api/llm/models: List available models

🤝 Contributing

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

📄 License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

Alternative licensing under Creative Commons CC BY-SA 4.0 International is also available.

🙏 Acknowledgments

• Built with Svelte 5 and FastAPI
• UI components styled with Tailwind CSS
• Markdown rendering by Marked
• Migrated to Svelte 5 runes API for improved performance and developer experience

📞 Support

• Documentation: https://spacecruft.org/deepcrayon/multibrain
• Issues: https://spacecruft.org/deepcrayon/multibrain/issues

---

Copyright © 2025 Jeff Moe