SQLite and FAISS

A high-speed vector search extension for SQLite that embeds FAISS capabilities directly into local SQL workflows.

The sqlite-vss extension (developed by Alex Garcia) brings production-grade vector similarity search to the SQLite ecosystem. By integrating the FAISS library, it allows developers to store and query high-dimensional embeddings (such as 768-dimension BERT or 1536-dimension OpenAI vectors) using standard SQL commands. The system utilizes a virtual table mechanism: users call the vss_search function to perform k-nearest neighbor lookups with millisecond latency. It is a lightweight, zero-config solution for building local RAG pipelines and recommendation engines without the complexity of a dedicated vector database.

https://github.com/asg017/sqlite-vss