VexDB-THU/VexDB-Lite

Health v2 · maintenance only

Flags: young

Adoption not part of the score

2275 stars · 353 forks observed · 2026-08-28

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

VexDB-Lite is a cross-platform vector similarity search engine implemented as plugins for PostgreSQL, DuckDB, and SQLite, sharing a graph index algorithm, SIMD distance dispatch, and PQ/RaBitQ quantization kernels. It provides vector types, distance operators, graph indexes with metadata filtering, and optimizer-integrated ANN index scans.

Use cases

• add vector similarity search to postgres
• semantic search over embeddings in sqlite
• nearest neighbor search in duckdb
• store and query embeddings alongside relational data
• run vector search on mobile or in the browser with sqlite/wasm
• compress vector indexes with pq or rabitq quantization

When to choose

• you want ANN search inside an existing PostgreSQL, DuckDB, or SQLite database rather than a separate vector database
• you need cross-platform support including mobile and WASM via SQLite
• you need quantized (PQ/RaBitQ) compact index storage to reduce memory

When to avoid

• you need a standalone distributed vector database with sharding and high availability - use the commercial VexDB or alternatives like Milvus
• you need BM25 hybrid full-text plus vector retrieval out of the box - the Lite plugins focus on vector search
• you require non-SQL databases such as MongoDB or Redis

Facets

library · maturity active

vector-database database search-engine database-driver databases machine-learning developer-tools windows wasm cpp cross-platform ann-search hnsw postgresql-extension duckdb-extension sqlite-extension quantization embeddings similarity-search search linux macos android ios

Member repositories

For agents

markdown · JSON · MCP: product_card(name="VexDB-THU/VexDB-Lite")