Juan is a friendly orange-and-white cat chatbot 🐱 designed to share engaging and insightful information about the world and his human, Fabio Guerreiro.

• Integrated into Fabio's portfolio, Juan helps visitors get to know him in a fun and interactive way.
• As a general-purpose chatbot at chatbot-juan.vercel.app showcasing the power of LLM and TTS

Deployed at: chatbot-juan.vercel.app

Powered by:

🧬 Next.js – the React framework for building and deploy fast and scalable web apps

Powered by:

• 🐍 Python – the core language used to build Juan
• 🧠 LangChain – to manage the reasoning and prompts
• 📚 RAG (Retrieval-Augmented Generation) – to provide smart, context-based answers
• 🐋 Docker – for easy deployment
• 🤖 Ollama – local LLM inference

• Serves a FastAPI endpoint for chat interactions
• Loads and processes contextual info using RAG
• Acts as a curious and cute AI cat named Juan
• Only answers what is asked, following a natural and concise tone

docker build -t juan .
docker run -d -p 8051:8051 juan
docker run juan

GET http://localhost:8051/status
POST http://localhost:8051/juan/chat

body:
  {
    "message": "Tell me about Fabio's projects"
  }

• Make sure you have Ollama running locally - (https://hub.docker.com/r/ollama/ollama)
• LangChain + RAG is used to fetch and ground responses using a knowledge base about Fabio.
• The chatbot try to avoid adding unnecessary info and responds directly based on user questions.

Open an issue or a PR if you want to contribute, or just say hi to Juan! 😸