zvec

zvec

Lightweight and ultra-fast in-process vector database.

Agent SkillDataOpen source向量数据库全文搜索混合检索
Type
Agent Skill
Open source
Yes
GitHub Stars
★ 12.3k
Source
skill-github

Overview

Zvec is an open-source in-process vector database that is lightweight and fast, allowing direct embedding into applications. Verified internally at Alibaba Group, it delivers production-grade low-latency and scalable similarity search capabilities without complex configuration. Zvec supports dense and sparse vectors, full-text search, and hybrid retrieval, and provides SDKs in multiple languages along with the visualization tool Zvec Studio. Ideal for applications requiring efficient vector search.

Capabilities

  • Millisecond-level search for billions of vectors
  • Supports dense and sparse vectors
  • Built-in full-text search functionality
  • Fusion of vector similarity and full-text search
  • Persistent storage to ensure data durability

Use cases

Large-scale vector similarity searchSemantic search in natural language processingUser behavior analysis in recommendation systemsSimilarity matching for image and video content

Setup

Requires: Python 3.10-3.14Node.jsGoRust
pip install zvec (Python) or npm install @zvec/zvec (Node.js)

This information was compiled by AI from public sources and may contain inaccuracies — please refer to the source.

FAQ

Which platforms does Zvec support?

Linux (x86_64, ARM64), macOS (ARM64), and Windows (x86_64).

How do I build Zvec from source?

Please refer to the build guide in the official documentation.

Related skills