Farfalle: Open-Source AI Search Engine 115

Farfalle.dev is an open-source AI-driven answer AI search engine, positioned as a self-hosted alternative to Perplexity. It allows users to run local large language models (LLMs), such as llama3, gemma, mistral, and phi3, or use custom LLMs through LiteLLM. In addition, it supports cloud models such as Groq/Llama3 and OpenAI's GPT-4 and GPT-3.5-turbo 2 4.

Key Features of Farfalle:

• Search Providers: Farfalle integrates with multiple search providers, including Tavily, Searxng, Serper, and Bing 2.

• Local and Cloud Models: Users can use local models (llama3, mistral, gemma, phi3) or cloud models (OpenAI/gpt4-o, OpenAI/gpt3.5-turbo, Groq/Llama3) to answer questions 2.

• Custom LLMs: Through LiteLLM, Farfalle supports custom LLMs, enhancing its flexibility 2.

• Tech Stack: The frontend is built with Next.js, and the backend is built with FastAPI. It also uses Redis for rate limiting and Logfire for logging 2.

• Deployment Options: Farfalle can be deployed using Docker or on platforms such as Render and Vercel 2 4.

Getting Started with Farfalle:

To get started with Farfalle locally:

Prerequisites: Ensure Docker and Ollama are installed if running local models.

API Keys: Obtain API keys for optional services such as Tavily, Serper, OpenAI, Bing, and Groq.

Run Command:

 docker run \
    -p 8000:8000 -p 3000:3000 -p 8080:8080 \
    --add-host=host.docker.internal:host-gateway \
    ghcr.io/rashadphz/farfalle:main

Visit http://localhost:3000 to access the application.

Farfalle source code is available on GitHub.https://github.com/rashadphz/farfalle