Nixiesearch is a hybrid search engine designed for consumer-facing applications, offering a modern alternative to complex Elasticsearch/OpenSearch clusters. It targets developers seeking a scalable, maintainable search solution with integrated semantic search capabilities, aiming to simplify operations and reduce infrastructure headaches.

How It Works

Nixiesearch leverages Apache Lucene for core search functionalities, including multi-language support, faceting, and advanced filtering. Its key innovation is a decoupled architecture where compute and storage are separated, with all index data residing on S3-compatible storage. This stateless design enables risk-free backups, upgrades, schema changes, and auto-scaling. It supports pull-based indexing via Spark ETL for both offline and online incremental updates, avoiding direct cluster manipulation.

Quick Start & Requirements

• Install/Run: docker run -itp 8080:8080 -v .:/data nixiesearch/nixiesearch:latest standalone -c /data/config.yml
• Prerequisites: Docker, S3-compatible storage, ONNX Runtime (for local embedding inference, SaaS providers also supported).
• Setup: Minimal setup time, primarily configuration.
• Links: Quickstart Guide, Live Demo

Highlighted Details

• Hybrid search (text + semantic) with integrated embedding inference and RAG API support.
• Decoupled S3-based storage and stateless compute for simplified operations and scalability.
• Apache Lucene foundation provides robust features like multi-language support, faceting, and suggestions.
• Immutable index configurations facilitate easy blue-green deployments.

Maintenance & Community

The project is open-source, inspired by search solutions from Uber, Amazon, and DoorDash. Community links and roadmap details are not explicitly provided in the README.

Licensing & Compatibility

• License: Apache 2.0.
• Compatibility: Permissive license suitable for commercial use and integration with closed-source applications.

Limitations & Caveats

Nixiesearch is explicitly not a database or a log search tool, focusing solely on relevance-based search for consumer applications. It is not designed for analytical workloads.