KarloAIMutthiMain

Our project is a fashion recommendation and virtual try-on system powered by LLMs and Couchbase Capella. We use a subset of 2,000 images from the Fashion Image Dataset, generating text embeddings from metadata to enable semantic vector search via Couchbase. Users input natural language queries to find matching clothing items, which are then used in a downstream style transfer task to overlay the garment onto a user-provided image. This creates a personalized and immersive shopping experience. The system leverages Nomic (Ollama) embeddings, Couchbase vector search, and generative AI models for virtual try-on.