# unum-cloud/USearch

Fast Open-Source Search & Clustering engine × for Vectors & Arbitrary Objects × in C++, C, Python, JavaScript, Rust, Java, Objective-C, Swift, C#, GoLang, and Wolfram 🔍

Repository: https://github.com/unum-cloud/USearch
Canonical: https://ross.abutalabs.com/products/usearch
Homepage: https://unum.cloud/usearch
Language: C++
License: Apache-2.0
License Family: permissive
Topics: kann, vector-search, approximate-nearest-neighbor-search, simd, search, similarity-search, database, faiss, search-engine, webassembly, clustering, nearest-neighbor-search, recommender-system, semantic-search, full-text-search, fuzzy-search, text-search, image-search
Last push: 2026-08-22T19:40:07+00:00

## Health v2 (maintenance only)
Score: 93/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 87, longevity 92
- inputs: {"age_days": 1288, "days_push": 11, "days_rel": 11, "gap_med": 1.0, "n_releases_24m": 75}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4278, forks 342 (observed 2026-08-28T04:08:41.069289+00:00)

## What it is
USearch is a fast, single-file similarity search and clustering engine for vectors and arbitrary objects, supporting spatial, binary, probabilistic, and user-defined metrics. It provides bindings for many languages including C++, Python, JavaScript, Rust, Java, Swift, C#, and Go, and runs on platforms from Linux to WebAssembly.

## Use cases
- build semantic search over document embeddings
- find nearest neighbors for image similarity search
- cluster high-dimensional vectors at billion scale
- add a lightweight vector index to an existing database like SQLite
- power a recommender system with approximate nearest neighbor search
- run vector search in the browser via WebAssembly
- search with custom distance metrics beyond cosine or euclidean

## When to choose
- you need a compact, embeddable ANN index with bindings for many languages
- you want faster or smaller alternative to FAISS for vector search
- you need custom distance metrics or non-vector object similarity
- you target unusual platforms like iOS, Android, or WebAssembly

## When to avoid
- you need a full distributed vector database with replication and persistence out of the box
- you only need exact k-NN on small datasets where brute force suffices
- you want managed cloud-hosted vector search

## Facets
- artifact type: library
- maturity: active
- function: search-engine, vector-database, database
- domain: machine-learning, databases, computer-vision
- platform: windows, wasm, cpp, python, rust, go, jvm, dotnet, cross-platform
- tags: ann, hnsw, similarity-search, embeddings, semantic-search, multi-language-bindings, single-file, simd, clustering, search, natural-language-processing, linux, macos, android, ios, swift

## Member repositories
- unum-cloud/USearch (main) score 93

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:41.069289+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 2026-08-29T18:21:59.445747+00:00, confidence not recorded.
  - readme: https://github.com/unum-cloud/USearch (fetched 2026-08-28T04:08:41.069289+00:00, sha 6ed3faf280da)
  - homepage: https://unum.cloud/usearch (fetched 2026-08-29T09:11:21.088589+00:00, sha 742bf1f7ecf7)
- Data as of 2026-08-30T08:39:29.467469+00:00.