# hora-search/hora

🚀  efficient approximate nearest neighbor search algorithm collections library written in Rust 🦀 .

Repository: https://github.com/hora-search/hora
Canonical: https://ross.abutalabs.com/products/hora
Homepage: http://horasearch.com/
Language: Rust
License: Apache-2.0
License Family: permissive
Topics: search-engine, rust, approximate-nearest-neighbor-search, artificial-intelligence, recommender-system, image-search, vector-search, algorithm, data-structures, simd, hnsw, similarity-search, neural-network, high-performance, machine-learning, k-nearest-neighbors, rust-sci, numeric
Last push: 2026-02-17T13:34:53+00:00

## Health v2 (maintenance only)
Score: 63/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 68, release rhythm 35, longevity 100
- inputs: {"age_days": 1936, "days_push": 197, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2658, forks 77 (observed 2026-08-28T04:07:07.395263+00:00)

## What it is
Hora is a Rust library of approximate nearest neighbor (ANN) search algorithms with SIMD acceleration and multithreading. It provides multiple index types (HNSW, SSG, PQIVF, BruteForce), multiple distance metrics, and bindings for Python, JavaScript, Java, and WebAssembly.

## Use cases
- build a vector similarity search for embeddings
- implement image search with nearest neighbors
- add a recommendation engine with k-NN lookup
- search high-dimensional vectors fast in Rust
- run approximate nearest neighbor search in Python
- deploy vector search in the browser via WebAssembly
- find similar face embeddings for face matching

## When to choose
- you need a lightweight, dependency-free ANN library in Rust without BLAS
- you want SIMD-accelerated HNSW, SSG, or PQIVF indexes
- you need portable vector search across desktop, mobile, and WebAssembly
- you want Python or JavaScript bindings for nearest neighbor search

## When to avoid
- you need a full vector database with persistence, filtering, and distributed deployment
- you require a managed service or REST API out of the box
- you need exact (non-approximate) nearest neighbor results at scale
- you need actively developed support for Go, Ruby, Swift, or R bindings

## Facets
- artifact type: library
- maturity: maintenance
- function: search-engine, machine-learning, math
- domain: machine-learning, artificial-intelligence
- platform: rust, python, cross-platform, wasm, windows
- tags: approximate-nearest-neighbor, ann, hnsw, vector-search, simd, similarity-search, k-nearest-neighbors, recommender-systems, image-search, search, algorithms, nodejs, linux, macos

## Member repositories
- hora-search/hora (main) score 63

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:07.395263+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 2026-08-30T02:18:26.081627+00:00, confidence not recorded.
  - readme: https://github.com/hora-search/hora (fetched 2026-08-28T04:07:07.395263+00:00, sha 51ea7f5bc05e)
  - homepage: http://horasearch.com/ (fetched 2026-08-29T10:01:18.533260+00:00, sha c9caed1ed7ad)
  - registry_crates: https://crates.io/api/v1/crates/hora (fetched 2026-08-29T10:01:18.542260+00:00, sha 142d9ab85624)
- Data as of 2026-08-30T08:39:29.467469+00:00.