Repository of Natural Language Processing project at Polytechnic of Milan. Generative chatbots, with audio, images and RAG.

The AudioChatBot App receives voice inputs and sends them to the backend, then waits for a response. Once a response is received, it plays the audio.

Showcasing the use of GPT-4o Audio model

A FastAPI-based Retrieval-Augmented Generation (RAG) chatbot with audio upload, Whisper transcription, FAISS vector search, and GPT-powered responses.