Is Zvec the ‘SQLite Moment’ for Vector Databases?

Zvec is an open‑source vector database that runs inside the application process rather than as a separate server.

Vector databases store and retrieve high‑dimensional vectors for similarity search, recommendation, and AI applications. Zvec builds on Alibaba’s internal Proxima engine and the official data indicate millisecond‑level search over tens of billions of vectors.

Installation consists of a single pip install zvec command; the library becomes usable within about 60 seconds.

Supported data types include dense and sparse vectors, and a single query call can perform mixed searches. Because it is an in‑process library, Zvec can run on laptops, servers, CLI tools, and edge devices. The project is released under the Apache 2.0 license.

import zvec

# Define collection schema
schema = zvec.CollectionSchema(
    name="example",
    vectors=zvec.VectorSchema("embedding", zvec.DataType.VECTOR_FP32, 4),
)

# Create collection
collection = zvec.create_and_open(path="./zvec_example", schema=schema)

# Insert documents
collection.insert([
    zvec.Doc(id="doc_1", vectors={"embedding": [0.1, 0.2, 0.3, 0.4]}),
    zvec.Doc(id="doc_2", vectors={"embedding": [0.2, 0.3, 0.4, 0.1]}),
])

# Vector similarity search
results = collection.query(
    zvec.VectorQuery("embedding", vector=[0.4, 0.3, 0.3, 0.1]),
    topk=10
)

Some developers note that embedding the vector‑database capability directly into applications can simplify query flow, which is advantageous for Retrieval‑Augmented Generation (RAG) systems or other local vector‑search scenarios.

Other observers raise concerns about memory consumption because vector data can be large, and a few users have claimed the presence of a security backdoor without providing concrete evidence.

Project repository and documentation are available at https://github.com/alibaba/zvec.