asg017/sqlite-vss

Health v2 · maintenance only

Adoption not part of the score

1998 stars · 75 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

sqlite-vss is a SQLite extension written in C++ that adds vector similarity search to SQLite using Faiss, via vss0 virtual tables and a vss_search function. It is no longer in active development; the author recommends its successor sqlite-vec.

Use cases

• add vector similarity search to an existing sqlite database
• build a semantic search engine with embeddings in sqlite
• store and query embeddings for a recommendation system
• run k-nearest-neighbor queries over embedding vectors locally
• build a lightweight RAG or Q&A tool without a separate vector database

When to choose

• you already use SQLite and want embedded vector search with no extra server
• you need a bring-your-own-embeddings local vector index based on Faiss
• you want a simple fts5-like API for nearest-neighbor queries

When to avoid

• starting a new project - use sqlite-vec instead, as sqlite-vss is not actively developed
• you need UPDATE support on vector rows
• you need frequent small inserts or deletes, which are slow
• you need a production-grade distributed vector database

Facets

library · maturity maintenance

vector-database search-engine database databases machine-learning windows cli sqlite-extension vector-search faiss embeddings knn semantic-search deprecated-in-favor-of-sqlite-vec search linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="asg017/sqlite-vss")