AI Potluck
Product / UX / Agent tools & protocols

Milvus

Zilliz

Cloud-native vector database built for scalable approximate nearest neighbor search at billion-scale. Supports hybrid search combining vector similarity with scalar filtering. The most-starred dedicated vector database on GitHub (44K stars) with strong enterprise adoption. Open-source core with Zilliz Cloud as the managed offering. The go-to choice for production RAG systems that need to scale.

v2.6.x stable (late 2025); v3.0 RC published May 2026; LF AI & Data project

Openness

5 high confidence
5.0
license
Apache-2.0
source
public
core-gated
ungated

license:Apache-2.0;source:public;LF-AI&Data-governed(Zilliz-Cloud is managed hosting, not gated core)

Adoption

3 high confidence
3.0

Described as the default open source engine for billion-scale semantic search; LF AI & Data graduate. No precise verified deployment/download count, rated 3 on reported traction.

Capability

4 high confidence
4.0

High-performance cloud-native vector ANN search; covered by independent ANN-Benchmarks and first-party VectorDBBench. v2.6 adds Storage Format V2, advanced JSON, BM25 full-text search. Rated on ANN benchmark coverage / scale.

Unchanged since 2026-07-30 (last edited, not re-checked)