Hi HN,

I built EdgeVec, a vector database that runs entirely in the browser. It implements HNSW (Hierarchical Navigable Small World) graphs for approximate nearest neighbor search.

Performance: - Sub-millisecond search at 100k vectors (768 dimensions, k=10) - 148 KB gzipped bundle - 3.6x memory reduction with scalar quantization

Use cases: browser extensions with semantic search, local-first apps, privacy-preserving RAG.

Technical: Written in Rust, compiled to WASM. Uses AVX2 SIMD on native, simd128 on WASM. IndexedDB for browser persistence.

npm: https://www.npmjs.com/package/edgevec GitHub: https://github.com/matte1782/edgevec

This is an alpha release. Main limitations: build time not optimized, no delete operations yet.

Would love feedback from the community!

Bringing HNSW graphs to the browser is an interesting approach. The 3.6x memory reduction through quantization is significant. Curious how this compares to other in-browser vector libraries in terms of search latency vs memory trade-offs.