This project solves the problem of providing accurate and reliable information about clinics by using a Retrieval-Augmented Generation pipeline. It works by transforming clinic information into vector embeddings, which are then used to generate responses to patient queries. The pipeline is designed to be multi-tenant, with each clinic having its own schema and allowlist. This approach ensures that the information provided is accurate and up-to-date, and that patients receive reliable answers to their questions. The project is built using a range of technologies, including OpenAI, Postgres, and Deno.
View on GitHub →iTristaoo/rag-knowledge-base