Fast semantic vector database (HNSW + mmap) for Node.js. Version 1.0.0 uses zero-copy memory-mapped storage for local-only deployment, with O(log n) approximate nearest neighbor search. Ships TypeScript types and prebuilt N-API binaries via @mmomtchev/prebuild-install, falling back to node-gyp rebuild (C++17, no CMake). Provides timestamp ranges, distance metrics (IP, Cosine, L2), and full CRUD. Supports Linux, macOS, Windows (x64, arm64). Node >= 16 required.
npm install logosdbNo compatibility data collected yet for this library.
Verified import paths — ran on the pinned version, not inferred.
Creates a database with cosine distance, inserts one vector, searches for top 5 nearest neighbors, and closes.
cd nodejs && npm publish
npm run native:prebuild
Preprocess text to embeddings before calling db.put().
Install Python 3.x and a C++17 compiler (gcc, clang, MSVC).
cd nodejs && npm publish
Ensure Python and C++17 toolchain are installed, or download prebuilt tarball from releases.
Use const { DB } = require('logosdb') or import { DB } from 'logosdb'