sqliteai/sqlite-vector

SQLite-Vector is a cross-platform, ultra-efficient SQLite extension that brings vector search capabilities to your embedded database.

What it solves

SQLite-Vector provides production-grade vector search capabilities directly within SQLite. It eliminates the need for external vector databases (like FAISS or Weaviate) or complex, time-consuming preindexing processes (like HNSW or DiskANN), making it ideal for Edge AI, mobile apps, and privacy-preserving offline workloads.

How it works

It is implemented as a cross-platform SQLite extension that allows vectors to be stored as standard BLOBs in ordinary tables. It uses highly optimized C implementation with SIMD acceleration for distance calculations. To improve performance and reduce memory usage, it employs various quantization techniques, including TurboQuant (a low-bit scalar quantizer) and other formats like INT8 or 1BIT, allowing the system to scan vectors without reconstructing them fully.

Who it’s for

Developers building RAG (Retrieval-Augmented Generation) or semantic search applications who want a lightweight, embedded solution that runs on iOS, Android, Windows, Linux, macOS, or in the browser via WASM.

Highlights

  • No preindexing: Start searching immediately without long preprocessing phases.
  • Ordinary tables: Store vectors as BLOBs without needing special virtual tables.
  • TurboQuant support: High-efficiency 2, 3, and 4-bit quantization scans.
  • Low memory footprint: Defaults to 30MB of RAM usage.
  • Cross-platform: Native support for all major operating systems and mobile platforms.
  • Offline ready: Works entirely on-device for privacy and speed.

Related

  • Project
    1yefuwang1/vectorlite

    Vectorlite is a Rust‑based SQLite extension that adds a virtual table for fast approximate nearest‑neighbor search on high‑dimensional vectors, using HNSW and SIMD‑accelerated distance calculations. It works on Linux, Windows and macOS Apple Silicon and is available via Python wheels and an npm package.

  • Project
    weaviate/weaviate

    A cloud-native vector database that stores objects and vectors to enable semantic search, hybrid search, and retrieval-augmented generation (RAG) at scale.

  • Project
    RyanCodrai/turbovec

    A Rust-based vector index with Python bindings that uses the TurboQuant algorithm to drastically reduce memory usage and increase search speed for vector search.

  • Project
    VexDB-THU/VexDB-Lite

    A high-performance vector similarity search engine for PostgreSQL, DuckDB, and SQLite that uses a custom graph index and quantization to enable efficient ANN search.

  • Project
    supervc-stack/VectorChord

    A PostgreSQL extension for high-performance, cost-effective vector search that enables billion-scale vector hosting on modest hardware using RaBitQ compression.