1. Project Overview

VectoRex

A high-performance vector search engine implemented entirely in Java.

VectoRex is a high-performance, scalable vector search engine designed for modern AI and big data applications. It offers lightweight, embeddable, and standalone deployment options, making it easy to handle complex scenarios such as recommendation systems, image search, and natural language processing. It provides developers with a powerful and flexible search solution.

2. Open Source License

Licensed under Apache-2.0

3. Screenshots

4. Feature Overview

Key Advantages

• Lightweight: No complex external libraries required. All core functionality is implemented in Java, making it easy to integrate and deploy with cross-platform compatibility. Truly write once, run anywhere.

• Embeddable: Provides starter and solon-plugin packages for immediate use. Simply add the dependency to quickly integrate it into existing systems and seamlessly fit into various application architectures.

• Standalone Deployment: Supports standalone search service deployment with a complete feature set. Clients can easily interact with the service through RESTful API interfaces to meet the needs of different use cases.

• SDK: Provides Java, Golang, and Python client SDKs for interacting with VectoRex services, offering developers a convenient integration experience.

Why Choose VectoRex?

Technical Features

• High-Performance Search: Powered by the advanced HNSW algorithm, it supports fast searches across large-scale vectors, providing users with an efficient Approximate Nearest Neighbor (ANN) search experience.

• Hybrid Queries: Seamlessly combines vector search with scalar filtering to support complex query scenarios, such as finding images most similar to a specific picture with the tag "landscape", or locating documents most similar to a given text that were published within the past week.

• Persistent Storage: Built-in efficient data storage mechanisms ensure secure storage and rapid recovery of large-scale datasets, maintaining data integrity and reliability.

• Flexible Index Management: Supports scalar indexes (inverted indexes and range indexes) and vector indexes, with dynamic addition, deletion, and updating of vector data to meet diverse indexing requirements.

Use Cases

• Recommendation Systems: Precisely identifies content that closely matches user interests, improving recommendation quality and user experience.

• Image Search: Uses image feature-based similarity search to quickly find image resources similar to a target image.

• Natural Language Processing: Enables semantic search and text similarity calculations, supporting efficient operation of natural language processing applications.

• Bioinformatics: Supports gene sequence alignment and protein structure search, providing powerful assistance for bioinformatics research.

5. Tech Stack

Lightweight Design

No complex external libraries required. All core functionality is implemented in Java, making integration and deployment easy.

Cross-Platform Compatibility

Write once, run anywhere. Compatible with multiple operating systems for seamless execution.

Easy Embedding and Integration

Provides starter and solon-plugin packages for immediate use and quick integration into existing systems.

Standalone Deployment Support

Supports standalone search service deployment with a complete feature set to meet diverse application requirements.

RESTful API Integration

Enables convenient client interaction through RESTful APIs to accommodate different development requirements.

Multi-Language SDKs