zvec
SHA-256Zvec is Alibaba's open-source, in-process vector database. Like SQLite for embeddings, it embeds directly into your app with no server to deploy, delivering millisecond similarity search, full-text search, and hybrid queries across Python, Node.js, Go, Rust, and Flutter.
Smart Download
Download Download Version
v0.8.0 · 39.4 MB
An embeddable, SQLite-style vector database for millisecond local similarity search.
Core Features
- Millisecond vector search scaling to billions of entries
- Fully embedded and local—no server, no configuration
- Dense, sparse, and multi-vector query support
- Native full-text search with structured expressions
- Hybrid queries fusing vectors, keywords, and filters
What It Can't Do
- •zvec is an embedded library, not a standalone server—it cannot serve remote clients over the network like Milvus or Qdrant. Writes are single-process exclusive; multiple processes may read concurrently but cannot write together. Data lives in local files, so backups and disaster recovery are your responsibility. Only 64-bit Python 3.10–3.14 is supported. Uniform quantization boosts speed but may reduce recall, so enable it carefully for precision-sensitive workloads.
Use Cases
- Local knowledge retrieval for RAG apps and AI agents
- Offline semantic search on mobile or edge devices
- Similarity search for documents, code, or images
- Real-time similar-item recall in recommender systems
Zvec is an open-source, in-process vector database developed and battle-tested at Alibaba Group, designed to embed directly into applications with minimal setup. Written in C++ with official SDKs for Python, Node.js, Go, Rust, and Dart/Flutter, it searches billions of vectors in milliseconds and runs on Linux, macOS, Windows, Android, and iOS. Unlike server-based vector databases such as Milvus or Qdrant, zvec needs no separate service—think SQLite rather than PostgreSQL—so it fits notebooks, CLI tools, edge devices, and mobile apps. Compared with index-only libraries like FAISS, zvec adds production essentials out of the box: write-ahead logging for crash-safe durability, document CRUD, dense and sparse vectors, native full-text search, and hybrid queries that fuse vector similarity, keywords, and structured filters in one call. Multi-process concurrent reads are supported, while writes are single-process exclusive. It is a strong fit for RAG pipelines, semantic search, recommendation, and local-first AI agents that need low-latency retrieval without operating a database cluster.
Tags
Getting Started
Download installer
Click the button above to download the installer for your system
Install the software
Double-click the downloaded installer and follow the prompts
Make sure you have 64-bit Python 3.10–3.14, then run pip install zvec
Define a CollectionSchema in your code with the collection name, vector field, and dimensions
Create and open a collection with create_and_open, then insert documents with vectors
- Make sure you have 64-bit Python 3.10–3.14, then run pip install zvec
- Define a CollectionSchema in your code with the collection name, vector field, and dimensions
- Create and open a collection with create_and_open, then insert documents with vectors
SHA-256 checksum verified
Checksum extracted from GitHub official Release page
SHA256 Checksum
98779b81a0c2c3845336b8f5da8203908116bbf55275805ae9124779e786dc20This checksum is extracted from the GitHub Release page. Verify file integrity after download.
All SHA-256 checksums on this platform are extracted from the project's official GitHub Release page, without any modification. You can independently verify them on the GitHub Releases page.
Open Source Transparency
View GitHub SourceUninstall Info
Run pip uninstall zvec to remove the Python package (or npm uninstall @zvec/zvec for Node.js). Then delete the local data directory (e.g., ./zvec_example) to remove all stored data—no background services or leftovers remain.
No Extra Dependencies
Ready to use after download. No additional runtime required.
Having issues? Check the FAQ below
4 FAQs
Similar Projects
Visual Studio Code (Code - OSS)
A free, open-source code editor by Microsoft with rich extensions, debugging, and Git integration. The most popular developer tool worldwide.
opencode
OpenCode is a free, open-source AI coding agent that helps you write, refactor, and understand code. It offers two built-in modes: a full-access 'build' agent for active coding and a read-only 'plan' agent for safe exploration.
scrcpy
Display and control your Android device from your computer via USB or TCP/IP. No root required, high performance, low latency, supports audio and video mirroring.