# sqliteai/sqlite-vector

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

Repository: https://github.com/sqliteai/sqlite-vector
Canonical: https://ross.abutalabs.com/products/sqlite-vector
Homepage: https://sqlite.ai
Language: C
License: NOASSERTION
License Family: other
Last push: 2026-08-25T08:11:13+00:00

## Health v2 (maintenance only)
Score: 86/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 99, longevity 36
- inputs: {"age_days": 513, "days_push": 8, "days_rel": 9, "gap_med": 1.5, "n_releases_24m": 53}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1100, forks 49 (observed 2026-08-28T04:03:34.905620+00:00)

## What it is
SQLite-Vector is a cross-platform SQLite extension written in C that adds vector search capabilities to embedded SQLite databases, with SIMD-accelerated distance functions and TurboQuant 2/3/4-bit quantization. It stores vectors as BLOBs in ordinary tables without virtual tables or preindexing, runs on mobile, browser, edge, and server, and uses only ~30MB of memory by default.

## Use cases
- add vector search to my sqlite database
- run semantic search offline on mobile devices
- store embeddings locally for a RAG app
- on-device similarity search with low memory footprint
- quantize vectors to 2/3/4-bit for edge AI
- build a local-first AI agent with embedding memory
- search vectors in the browser with sqlite wasm

## When to choose
- you need vector search inside an existing SQLite app without a separate vector database
- you're building offline-first or on-device AI with tight memory constraints
- you want quantized vector scans (TurboQuant) with SIMD acceleration
- you need cross-platform support including mobile and browser (WASM)

## When to avoid
- you need distributed, multi-node vector search at massive scale
- you require managed infrastructure with built-in auth and sync (consider SQLite Cloud or a dedicated vector DB)
- your workload needs ANN indexes over billions of vectors rather than exact scans

## Facets
- artifact type: library
- maturity: active
- function: vector-database, database, search-engine, sdk
- domain: databases, machine-learning, artificial-intelligence, developer-tools, embedded-systems
- platform: windows, cross-platform, wasm, browser
- tags: sqlite-extension, vector-search, embeddings, turboquant, quantization, simd, edge-ai, on-device-ai, rag, ios, android, linux, macos

## Member repositories
- sqliteai/sqlite-vector (main) score 86

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:34.905620+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T06:45:55.848971+00:00, confidence not recorded.
  - readme: https://github.com/sqliteai/sqlite-vector (fetched 2026-08-28T04:03:34.905620+00:00, sha 8409bd7cb950)
  - homepage: https://sqlite.ai (fetched 2026-08-29T12:48:59.293528+00:00, sha 99d8968ff13d)
  - site_page: https://www.sqlite.ai/pricing (fetched 2026-08-29T12:48:59.303441+00:00, sha f02c43a2a836)
  - site_page: https://www.sqlite.ai/changelog (fetched 2026-08-29T12:48:59.305679+00:00, sha 30a18014da99)
- Data as of 2026-08-30T08:39:29.467469+00:00.