Quorum - Multi-LLM Conversation Platform

Quorum enables simultaneous conversations with multiple Large Language Models, allowing you to compare responses, synthesize insights, and leverage the strengths of different AI providers in a single interface.

Note: Quorum is designed to be self-hosted — you bring your own API keys and run it locally or on a server only you can reach. The /api/llm endpoint has no auth or rate-limiting, so don't expose a Quorum instance to the public internet without adding those yourself.

Features

• 🤖 Multi-Model Conversations - Query multiple LLMs simultaneously
• 🔄 Response Synthesis - AI-powered synthesis of multiple model responses
• 💰 Cost Tracking - Real-time cost estimation for API usage
• 🎨 Intuitive UI - Clean, responsive interface with model-specific color coding
• ⚡ Latest Models - Support for cutting-edge models from all major providers
• 🎛️ Customizable - Adjustable temperature and system prompts

Supported Models

Quorum currently ships configs for 80+ models across 10+ providers. A few highlights:

• Anthropic — Claude Fable 5, Mythos 5, Opus 4.8, Sonnet 4.6, Haiku 4.5
• OpenAI — GPT-5.5, GPT-5.4 (+ Mini/Nano), o3, o4-mini
• Google — Gemini 3.1 Pro, 3.5 Flash, 2.5 Pro/Flash
• xAI — Grok 4.3, Grok 4.20 Beta, Grok 4
• Meta — Llama 4 Maverick, Llama 4 Scout (via DeepInfra)
• DeepSeek — V4 Pro/Flash, V3.2 Exp, R1 (via DeepInfra)
• Qwen — Qwen3.5 / Qwen3.6 family (via DeepInfra)
• Moonshot — Kimi K2.6, K2.5 (via DeepInfra)
• Zhipu AI — GLM-5.1, GLM-4.7 (via DeepInfra)
• Open source — Gemma, Mistral, GPT-OSS (via DeepInfra)

The full registry — including pricing, context windows, and provider routing — lives in lib/llm.js. To add or update a model, edit modelConfigs there.

Prerequisites

• Node.js 18+
• npm or yarn
• API keys for the providers you want to use

Installation

1. Clone the repository:

git clone https://github.com/tap2k/quorum.git
cd quorum

2. Install dependencies:

npm install

3. Set up environment variables:

cp .env.example .env

4. Add your API keys to .env. You only need keys for the providers whose models you actually plan to use:

ANTHROPIC_API_KEY=your_anthropic_key
OPENAI_API_KEY=your_openai_key
GOOGLE_API_KEY=your_google_key
XAI_API_KEY=your_xai_key
DEEPINFRA_API_KEY=your_deepinfra_key   # gateway for Meta, DeepSeek, Qwen, Moonshot, Zhipu, Gemma, Mistral, GPT-OSS

# Optional
NEXT_PUBLIC_SITE_URL=http://localhost:3000
DEBUG_LLM=1                   # log reasoning/thinking output to server console

5. Start the development server:

npm run dev

6. Open http://localhost:3000

Usage

Basic Conversation

1. Select models by clicking on them in the model bar
2. Type your message and press Enter
3. View responses from all selected models side-by-side
4. Compare different perspectives and approaches

Response Synthesis

• Click "Synthesize" after receiving multiple responses
• An AI model will analyze and combine insights from all responses
• Useful for getting a balanced, comprehensive answer

Customization

• System Prompt: Click "System Prompt" to set custom instructions
• Temperature: Adjust creativity/randomness (0 = focused, 1 = creative)
• Synthesis Model: Choose which model synthesizes responses

Cost Tracking

• Individual response costs shown next to model names
• Total conversation cost displayed in header
• Costs estimated based on token usage

API Endpoints

/api/llm

Handles all LLM operations:

• action: 'chat' - Single model conversation
• action: 'multi-chat' - Multiple models in parallel
• action: 'synthesize' - Synthesize multiple responses

Architecture

quorum/
├── pages/
│   ├── index.js              # Main UI and conversation logic
│   └── api/
│       └── llm.js            # API endpoint for LLM calls
├── lib/
│   └── llm.js                # LLM service and model registry
├── styles/
│   └── globals.css           # Tailwind CSS styles
├── public/                   # Static assets
├── next.config.mjs
├── eslint.config.mjs
├── postcss.config.mjs
├── jsconfig.json
├── package.json
├── AGENTS.md                 # Codebase guide (also linked from CLAUDE.md)
├── LICENSE
└── README.md

Configuration

Adding New Models

Edit lib/llm.js and add to modelConfigs:

'model-id': {
  provider: 'provider-name',
  modelName: 'api-model-name',
  displayName: 'Display Name',
  color: '#HexColor',
  cost: { input: 0.00, output: 0.00 }, // per million tokens
}

Temperature Handling

Reasoning models (OpenAI o-series, GPT-5+, Claude Fable 5, Mythos 5, Opus 4.8/4.7, and others) don't accept a custom temperature and instead use a reasoning_effort parameter. lib/llm.js handles these per-model differences automatically — you can leave the temperature slider in the UI and the right thing will happen.

Development

# Run development server
npm run dev

# Build for production
npm run build

# Start production server
npm run start

# Run linter
npm run lint

Troubleshooting

Model Errors

• "Field required": Check if the model expects different parameter names
• "Temperature not supported": Some models only support default temperature
• "API key not configured": Ensure the corresponding API key is set in .env

Cost Calculations

Costs are estimated using character count (1 token ≈ 4 characters). For exact costs, refer to provider documentation.

Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Test thoroughly
5. Submit a pull request

License

MIT License - see LICENSE file for details

Acknowledgments

Built with:

• Next.js - React framework
• Tailwind CSS - Styling
• Provider APIs: Anthropic, OpenAI, Google, xAI, DeepInfra

Support

For issues, questions, or suggestions, please open an issue on GitHub.