Vector Database Comparison Cheat Sheet 2026: 8 Databases at a Glance

One-page cheat sheet comparing 8 vector databases by stars, license, and best use case.

## Vector Database Comparison Cheat Sheet 2026 All star counts from the GitHub API snapshot on 2026-07-31. Use this as a one-page decision aid. ### The 8 databases | Database | Stars | License | Best for | |----------|-------|---------|----------| | Milvus | 45,439 | Apache-2.0 | Enterprise scale, billions of vectors | | Qdrant | 33,697 | Apache-2.0 | Production Rust service, filters | | Weaviate | 16,677 | BSD-3 | Graph + vector hybrid, out-of-box modules | | pgvector | 22,417 | PostgreSQL | Adding vectors to existing Postgres | | Chroma | 30k+ | Apache-2.0 | Local dev, quick prototyping | | Typesense | 26,385 | GPL/Commercial | Typo-tolerant keyword + vector search | | Elasticsearch (kNN) | 40k+ | Elastic License | Already on Elastic stack | | Vespa | 7,034 | Apache-2.0 | Real-time ranking at scale | ### Decision rules 1. **Already use Postgres?** pgvector - zero new infrastructure 2. **Building a serious product?** Qdrant or Milvus 3. **Prototyping on a laptop?** Chroma - pip install chromadb 4. **Need hybrid search (keyword + vector)?** Typesense or Weaviate 5. **Billions of vectors + enterprise support?** Milvus (Zilliz behind it) ### Quick setup comparison - Chroma: pip install chromadb, 3 lines to embed and query - pgvector: CREATE EXTENSION vector; one column type - Qdrant: Docker run -p 6333:6333 qdrant/qdrant - Milvus: docker compose up -d (heavier) ### FAQ **Which is fastest?** At single-node scale they are similar; at 10M+ vectors Milvus and Qdrant pull ahead. **Do I need a vector database at all?** Under ~1M vectors, Postgres + pgvector or even SQLite with an embedding column is enough.