OPEN-SOURCE · 2026 EDITION

Vector DB
Showdown

A no-fluff comparison of the five most-used open-source vector databases. Sort any column. Read the chart. Pick your weapon.

Chroma Qdrant Weaviate Milvus pgvector
5 Databases
7 Dimensions
~80k Combined GitHub Stars
100% Open Source

Comparison Matrix

Click any column header to sort ↕

Database ↕ Licence ↕ Core Language ↕ Hosting Modes ↕ Index Types ↕ Filtering ↕ GitHub Stars ↕

GitHub Stars at a Glance

Community size as a proxy for ecosystem maturity (mid-2026)

Quick Verdict

Best for Prototyping
Chroma

Zero-config, Python-native, in-process. From pip install to first query in under a minute.

Best Filtering Performance
Qdrant

Rust engine with one-stage filtered HNSW. Payload filters are first-class, not an afterthought.

Best Hybrid Search
Weaviate

Native BM25 + vector fusion, built-in embedding modules, GraphQL and REST APIs out of the box.

Best for Billion-Scale
Milvus

Distributed architecture, the widest index portfolio (HNSW, IVF, SCANN, PQ), and the largest community.

Best for Existing Postgres
pgvector

Full SQL WHERE filtering, ACID transactions, and zero new infrastructure if you already run Postgres.