The current chatbot is stateless or uses local memory. Implement a PostgreSQL vector database (using Supabase pgvector) to store chat session histories so users can resume conversations, and use embeddings to retrieve domain-specific FreshScan docs to answer questions.
The current chatbot is stateless or uses local memory. Implement a PostgreSQL vector database (using Supabase pgvector) to store chat session histories so users can resume conversations, and use embeddings to retrieve domain-specific FreshScan docs to answer questions.