OpenSearch
OpenSearch is an open-source distributed, RESTful search engine. Its README calls it an “open-source, enterprise-grade search and observability suite that brings order to unstructured data at scale”. The site describes it as vendor-neutral, Apache 2.0 licensed “from ingest to dashboard”, and a Linux Foundation project run by the OpenSearch Software Foundation, formed in September 2024.
Repository snapshot (GitHub API, 2026-11-19)
| Item | Detail |
|---|---|
| Repo | opensearch-project/OpenSearch, Apache-2.0 |
| Created | 2021-01-29 |
| Activity | last push 2026-11-19; latest release 3.10.0 (2026-11-18) |
| Size | 14,035 stars, 3,089 forks |
Capabilities (opensearch.org)
- Search: enterprise and document search.
- AI and ML: vector search and anomaly detection.
- Observability: an observability stack and log analysis.
- Security analytics: threat intelligence and event correlation.
Where it fits
For RAG, OpenSearch is a search engine that stores both text (keyword relevance) and vectors, so one system can run keyword, vector and hybrid retrieval; see hybrid-search-and-rank-fusion. It is the open-source alternative to Elasticsearch and the managed counterpart to azure-ai-search; unlike faiss, it is a full server with persistence, filtering and an API. See vector-databases-compared for the dedicated databases.
Sources
- opensearch.org, the OpenSearch README and GitHub API (2026-11-19). The pages read do not describe k-NN or hybrid-search features in detail, nor the project’s relationship to Elasticsearch or AWS’s managed service, so those are not recorded here.