MoA Chat - Documentation

📖 What is MoA Chat?

MoA Chat is a simple but powerful chat platform where multiple AI models answer the same question at the same time, and an aggregator model combines their outputs into one final answer.

• Built in Python 3.
• Web framework: Flask.
• Frontend: HTML, JavaScript, TailwindCSS (optionally removable).
• Designed to work first on Hugging Face Spaces, but can also be self-hosted.

⚙️ Features

• Send your question once — multiple AI models answer simultaneously.
• Aggregator model (LLM-D) summarizes all responses.
• Fully configurable: choose which models you want to use.
• Modern minimal UI with light/dark theme toggle.
• Spanish/English documentation switch.
• Free models supported through OpenRouter and others.
• No API keys exposed in the frontend (safe backend request).

🛠️ Self Hosting

You can clone the project like this (includes the Dockerfile for containerization):

git clone https://huggingface.co/spaces/UntilDot/Flask

Requirements:

• Python 3.11+
• Pip
• Create a .env file and add your API keys.

Install dependencies:

pip install -r requirements.txt

Run locally:

python app.py

Default port is 7860 (to match Hugging Face standard).

🐳 Docker support:

Your repository includes a Dockerfile for easy containerization when you clone it. Here's the content:

# Use a slim Python base
FROM python:3.11-slim

# Set working directory
WORKDIR /app

# Install dependencies
COPY requirements.txt ./
RUN pip install --no-cache-dir -r requirements.txt

# Copy source code
COPY . .

# Expose the default Hugging Face Spaces port
EXPOSE 7860

# Run the app
CMD ["python", "app.py"]

Alternatively, you can manually create your own Dockerfile with the above content if you prefer to customize it.

To build and run the Docker container after cloning the repository:

docker build -t moa-chat .
docker run -d -p 7860:7860 --env-file .env moa-chat

Docker will NOT automatically inject secrets unless you:

• Use a .env file with --env-file .env
• Manually use -e VAR=VALUE flags in docker run

🔑 Environment Variables (Secrets)

To use this app, you must set your API keys in secrets or environment variables.

Follow this syntax:

OPENROUTER_API_KEY=your-openrouter-key
TOGETHER_API_KEY=your-together-key
GROK_API_KEY=your-grok-key
GROQ_API_KEY=your-groq-key

You can set these in:

• Hugging Face Secrets section (recommended if on Spaces)
• .env file (only for self-hosting)

🧩 How to Add More Models

All models and providers are declared inside:

llm/model_config.json

The structure looks like this:

{
  "providers": {
    "openrouter": {
      "url": "https://openrouter.ai/api/v1/chat/completions",
      "key_env": "OPENROUTER_API_KEY"
    }
  },
  "models": {
    "deepseek/deepseek-chat-v3-0324:free": "openrouter"
  }
}

To add a new model:

• Find the right provider (OpenRouter, Together, Grok, Groq, etc).
• Add its endpoint URL under "providers" if not already listed.
• Add your model name under "models" section, linking it to the provider.

Make sure your environment variables (secrets) are correctly configured.

🏷️ Licensing

This project is licensed under Apache 2.0 — You are free to use, modify, and distribute, even commercially.