Case Study: Quivr scales to 5,100 vector databases with Supabase

How one of the most popular Generative AI projects uses Supabase as their Vector Store

Quivr is an open source “second brain” app that lets users upload documents, audio, URLs, text, and code, then search and ask questions with generative AI. As the project grew from a weekend prototype into a viral AI product, Quivr needed a scalable place to store and query millions of embeddings, so it chose Supabase Vector and Supabase Auth.

Supabase provided the vector store, authentication, and supporting platform features that helped Quivr keep the product simple while scaling. With Supabase Vector powered by pgvector, Quivr is storing over 1.6 million embeddings and supporting similarity search across 100,000+ files; the open source community also helped drive growth to 5,100 Quivr databases on Supabase, 17,000 hosted signups in just over two months, and 500 daily active users.

Quivr

Stan Girard

Founder