Nixiesearch

Open source

Nixiesearch is a search engine built on Apache Lucene that keeps its index on S3-compatible storage and includes local embedding and LLM inference.

Open-source alternative to

nixiesearch.ai
Nixiesearch homepage screenshot
GitHub stars
620
Last 30 days
+1(+0%)
Last commit
10 mo ago
Repository age
3 years
Version
0.8.0
Licence
Apache-2.0
Self-hosted
Yes

About Nixiesearch

Nixiesearch is a search engine that runs on S3-compatible storage. Its authors built it after struggling with large Elastic and OpenSearch clusters. It is powered by Apache Lucene, with support for 39 languages, facets, advanced filters, autocomplete suggestions and sorting.

Storage and compute are decoupled. The index is stateless and kept in S3, and all changes to settings and indexes are config updates, which the project says makes backups, upgrades, schema changes and auto-scaling straightforward. It can run on Kubernetes, in Docker, or serverless with AWS Lambda.

Indexing is pull-based, supporting offline and online incremental indexing through an Apache Spark based ETL process. Embedding and LLM inference run locally, and there is first-class RAG API support. The docs also state what it is not: a database, or a tool for log search, where throughput matters more than relevance.

Key features

  • Index stored on S3-compatible storage
  • Lucene-based facets, filters and autocomplete
  • Runs on Kubernetes, Docker or AWS Lambda
  • Pull indexing with a Spark-based ETL process
  • Local embedding and LLM inference with RAG API
  • Settings and index changes are config updates

Good fit for

  • Site and product search
  • Semantic and hybrid search with embeddings
  • RAG over indexed documents
Built with
Scala tools
Tags
search engine
lucene
s3
semantic search
rag
kubernetes
embeddings

Nixiesearch: questions and answers

What is Nixiesearch used for?
Nixiesearch is a search engine built on Apache Lucene that keeps its index on S3-compatible storage and includes local embedding and LLM inference. It is a good fit for site and product search, semantic and hybrid search with embeddings, and RAG over indexed documents.
Is Nixiesearch open source?
Yes. Nixiesearch is open source under the Apache-2.0 licence. Its source code is on GitHub at nixiesearch/nixiesearch and is written mainly in Scala.
Is Nixiesearch free?
Yes. Nixiesearch is open source, so the software itself is free to use.
Can I self-host Nixiesearch?
Yes. Nixiesearch can be self-hosted on your own server or infrastructure.
What is Nixiesearch an alternative to?
Nixiesearch is an open-source alternative to Algolia, Elastic, Azure AI Search and Vectara. Other open-source alternatives to Algolia include Apache Solr, Vespa and Typesense.
Is Nixiesearch actively maintained?
The most recent commit to Nixiesearch was on 6 January 2026, and the latest release is 0.8.0, published on 6 January 2026. The project has 620 stars on GitHub, with 1 added in the last 30 days.

Open-source alternatives to Nixiesearch

See all

SaaS alternatives to Nixiesearch

See all