you have two main options for getting local models:

# Install Ollama from https://ollama.ai # Then pull any model: ollama pull llama3.2:1b # Llama 3.2 ollama pull mistral # Mistral 7B 

1. visit Hugging Face
2. search for compatible models (e.g., Llama 3)
3. download using Git LFS:

# Example for Mistral 7B git lfs install git clone https://huggingface.co/mistralai/Mistral-7B-v0.1

choose one of these servers:

• ollama

  ollama serve

• mlx omni server (for mac with apple silicon)

  pip install mlx-omni-server mlx-omni-server

• litellm

make your local server accessible using ngrok or localtunnel:

# For Ollama ngrok http 11434 --host-header="localhost:11434" # For MLX Omni Server ngrok http 10240

1. go to web.fullmoon.app
2. open settings
3. enter your endpoint details:
   • endpoint URL: https://your-ngrok-url.ngrok.io/v1
   • model name: Same as the model you downloaded (e.g., llama2, mistral)

a monorepo containing two versions of the fullmoon web app

1. a local version using SQLite for storage
2. a web version using IndexedDB for client-side storage live on https://web.fullmoon.app

apps/ ├── local/ # SQLite-based version └── web/ # IndexedDB-based version packages/ ├── database/ # Shared database interface └── ui/ # Shared UI components 

• Node.js 18+
• pnpm 8+

pnpm install

For local version (sqlite):

# Setup database npx prisma migrate dev # Start development server pnpm dev --filter local

For web version (IndexedDB):

pnpm dev --filter web

# Build all pnpm build # Build specific app pnpm build --filter local # or pnpm build --filter web

MIT