Nixiesearch is a hybrid search engine that fine-tunes to your data.
- Designed to be cloud-native with S3-compatible index persistence. Distributed with stateless searchers and scale-to-zero. No more
status: red
on your cluster. - Built on top of battle-tested Apache Lucene library: 39 languages, facets, advanced filters, autocomplete suggestions and sorting out of the box.
- Batteries included: RAG queries and vector search within a single container with a fully local CPU and GPU inference.
- Can learn the intent of a visitor by fine-tuning an embedding model to your data. Is "ketchup" relevant to a "tomato" query? It depends, but Nixiesearch can predict that from past user behavior.
Want to learn more? Go straight to the quickstart and check out the live demo.