// rag · alibaba/zvec

zvec

zvec é um projeto de RAG e conhecimento de código aberto de alibaba (alibaba/zvec no GitHub). Tem 16.079 estrelas no GitHub e 1.008 forks, é publicado sob a licença Apache-2.0 e foi atualizado pela última vez no GitHub em 29 de setembro de 2026.

Lightweight, lightning-fast in-process vector database from Alibaba that embeds in your app.

★ 16,1 mil estrelas⑂ 1,0 mil forksApache-2.0atualizado set 2026

Ver no GitHub ↗Official site ↗

▦ visual: o README deste projeto não inclui nenhuma captura incorporável,
então aqui vai um marcador. O guia abaixo está completo mesmo assim.

// dados do projeto

Dados do projeto

Repositório: alibaba/zvec
Estrelas no GitHub: 16.079
Forks: 1.008
Licença: Apache-2.0
Última atualização no GitHub: 29 de setembro de 2026
Categoria: RAG e conhecimento
Tipo de ficha: Guia de instalação completo
Site oficial: Official site

// como funciona

Como funciona

Decide where search should live
Choose zvec when you want vector search inside your process - no separate database server to deploy, configure or keep running.
Install the SDK for your language
Add the Python, Node.js, Go, Rust or Dart package to your project in one command.
Define a collection and insert vectors
Create a schema for your collection, insert documents with their embeddings, and let zvec build the index.
Query with hybrid search
Run vector similarity, full-text keyword or combined hybrid queries, and tune index and quantization options as data grows.

// recursos

Por que as pessoas gostam tanto

In-process embedding

Zvec runs inside your application with no server, no config and no fuss - install it and start searching in seconds.

Dense and sparse vectors

Supports dense and sparse embeddings, multi-vector queries, and a rich selection of vector index types that scale from memory to disk.

Full-text search

Native keyword-based full-text search lets you query string fields with natural-language or structured expressions.

Hybrid search

Combines vector similarity and full-text results in one system for retrieval that catches both meaning and exact terms.

Multiple language SDKs

Official packages for Python (pip), Node.js (npm), Go, Rust and Dart/Flutter, all backed by the same C++ core.

Quantization and disk indexes

Index options including IVF-RaBitQ and PQ-INT8 quantization, plus DiskANN support, keep large collections fast and compact.

// execute você mesmo

Execute você mesmo

Clone the repository
Clone the zvec source if you want to build from source or explore the C++ core and examples.
```
git clone https://github.com/alibaba/zvec.git
cd zvec
```
Install the Python SDK
Most users simply install the prebuilt package - 64-bit Python 3.10 to 3.14 - rather than building from source.
```
pip install zvec
```
Run the one-minute example
Create a collection, insert vectors and query them using the quickstart example from the README and docs.
```
python -c "import zvec; print(zvec.__version__)"  # then follow the one-minute example - see README for the exact code
```

// casos de uso

O que você pode criar

Adding semantic search to a desktop or mobile app without running a separate vector server.
Storing LLM and agent memory locally with fast HNSW-style similarity lookup.
Building RAG retrieval over documents with hybrid vector plus keyword search.
Powering on-device or edge search where a hosted vector database is not an option.

// perguntas frequentes

Perguntas frequentes

Who maintains zvec?

Zvec is an Alibaba open-source project, battle-tested within Alibaba Group, with community contributors.

Is zvec free to use?

Yes. It is Apache-2.0 licensed, so you can use it in commercial products.

Does zvec need a paid API key?

No. It runs locally in your process; you only need your own embedding model or API to turn text into vectors.

How does zvec compare with FAISS or Milvus?

FAISS is a library for vector search only, and Milvus is a server database; zvec is an embeddable database that adds full-text and hybrid search, schemas and multiple language SDKs on top of fast vector indexes.

How do I get started?

Os outros projetos de RAG e conhecimento com mais estrelas no catálogo do LibreAI — a lista realista de alternativas se zvec não for o ideal para você.

MarkItDown★ 193,5 mil
MarkItDown — Microsoft tool that converts PDFs, Office files, images and more to clean Markdown for LLMs, MIT
Dify★ 157,8 mil
Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or…
ragflow★ 91,6 mil
RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context…
PaddleOCR★ 90,8 mil
Lightweight OCR toolkit turning PDFs and images into structured, LLM-ready data in 100+ languages
MinerU★ 82,5 mil
MinerU — turns PDFs and Office documents into LLM-ready Markdown and JSON for agents and RAG pipelines, Apache-2.0 plus terms

// mais em rag

Mais em RAG e conhecimento

Projetos de RAG e conhecimento próximos no catálogo, os mais próximos em estrelas de zvec: continue explorando dos dois lados dele.

SurfSense★ 16,3 mil
Air-gapped, open-source NotebookLM alternative that turns local documents into briefings, decks and podcasts.
llmware★ 14,8 mil
Unified framework for enterprise RAG pipelines with small, specialized models - local, private and secure.
turbovec★ 17,4 mil
A blazing-fast vector index in Rust, built on TurboQuant - with Python bindings
QAnything★ 14,2 mil
Local knowledge-base Q&A over PDFs, Office files, images and web pages - offline and CPU-friendly.
DocsGPT★ 18,3 mil
A private AI platform for agents, assistants and enterprise search over your documents
LEANN★ 13,0 mil
The smallest vector index in the world: private RAG over everything on your laptop with 97% less storage.

Baixe o zvec

Grátis e aberto: pegue o código-fonte, dê uma estrela no repositório e comece a criar.

Ver no GitHub ↗← Mais ferramentas de RAG e conhecimento