What vector DB is best for search plus metadata filtering?

If your priority is both vector search and strong metadata filtering, the “best” choice depends on scale and how much operational complexity you want to handle. A few strong options:

Best overall for most teams: Pinecone

• Very good metadata filtering
• Managed service, easy to use
• Good latency and scaling
• Great choice if you want to avoid infrastructure work

Best open-source / self-hosted option: Qdrant

• Excellent payload/metadata filtering
• Strong performance
• Easy to operate compared with many self-hosted alternatives
• Good if you want control and lower vendor lock-in

Best if you already use PostgreSQL: pgvector + Postgres

• Works well for smaller to medium workloads
• SQL metadata filtering is a big advantage
• Simple stack if you already have Postgres
• Usually not the best for very large/high-QPS vector workloads

Best for enterprise search at large scale: Weaviate

• Supports vector + metadata filtering well
• Good schema/modeling features
• Can be great for search-heavy applications
• More complex than Pinecone/Qdrant in some deployments

Best if you need heavy analytics + search infrastructure: Milvus

• Strong at scale
• Good performance for large vector datasets
• Metadata filtering exists, but operational complexity can be higher

Quick recommendation

• Want easiest production path? → Pinecone
• Want open-source and strong filtering? → Qdrant
• Already on Postgres? → pgvector
• Large enterprise search platform? → Weaviate or Milvus

If you tell me:

1. dataset size
2. QPS / latency target
3. whether you need self-hosting
4. what metadata filtering looks like

…I can recommend a specific DB more confidently.