Weaviate.
Open-source vector database with strong hybrid search — Amsterdam-built.
What it is.
Weaviate is an open-source vector database written in Go. Bob van Luijt and the founder founded the company in Amsterdam in 2019 and pioneered the hybrid-search pattern that combines dense embeddings with sparse BM25. The platform ships as managed cloud, Bring Your Own Cloud, and a self-hosted Apache 2.0 release. Index Ventures led the $50 million Series B in 2023.
Where it fits.
Teams that want hybrid search without writing it twice. Particularly strong in European enterprises where data residency requires EU-hosted infrastructure. The open-source release ships the full feature set — paid tiers add multi-tenancy, dedicated nodes, and enterprise support rather than locking core capability.
- Hybrid search built in — dense plus BM25 in one query
- Apache 2.0 license — full self-host without feature gates
- Amsterdam-headquartered with EU data residency
- Operational overhead higher than Pinecone managed tier
- Java client lags Python and Go in feature coverage
Frequently asked.
Weaviate or Pinecone?
Weaviate when hybrid search or self-hosting matters. Pinecone when managed simplicity and US-cloud presence matter more.
Is Weaviate production-ready?
Yes. Used at scale by Stack Overflow, Instabase, and Morningstar. The Apache 2.0 release runs in production at companies that cannot use managed cloud.
Does Weaviate support multi-modal data?
Yes. Native modules handle text, image, and audio embeddings. CLIP and similar models work without external orchestration.
