open-source answer engine - run local or cloud models

Open-source AI-powered search engine. Run your own local LLM or use the cloud.

Demo answering questions with llama3 on my M1 Macbook Pro:

farfalle.dev (Cloud models only)

• 🛠️ Tech Stack
• 🏃🏿‍♂️ Getting Started
• 🚀 Deploy

• Add support for local LLMs through Ollama
• Docker deployment setup

• Frontend: Next.js
• Backend: FastAPI
• Search API: Tavily
• Logging: Logfire
• Rate Limiting: Redis
• Components: shadcn/ui

• Docker
• Ollama
  • Download any of the supported models: llama3, mistral, gemma
  • Start ollama server ollama serve

• Tavily
• OpenAI (Optional)
• Groq (Optional)

git clone git@github.com:rashadphz/farfalle.git
cd farfalle

touch .env

Add the following variables to the .env file:

TAVILY_API_KEY=...

# Cloud Models
OPENAI_API_KEY=...
GROQ_API_KEY=...

# Rate Limit
RATE_LIMIT_ENABLED=
REDIS_URL=...

# Logging
LOGFIRE_TOKEN=...

# API URL
NEXT_PUBLIC_API_URL=http://localhost:8000

# Local Models
NEXT_PUBLIC_LOCAL_MODE_ENABLED=true
ENABLE_LOCAL_MODELS=True

This requires Docker Compose version 2.22.0 or later.

docker-compose -f docker-compose.dev.yaml up -d

Visit http://localhost:3000 to view the app.

After the backend is deployed, copy the web service URL to your clipboard. It should look something like: https://some-service-name.onrender.com.

Use the copied backend URL in the NEXT_PUBLIC_API_URL environment variable when deploying with Vercel.

And you're done! 🥳