Weaviate
weaviate.io · Vector Databases
Weaviate is an open source vector database in Go with hybrid BM25 plus vector search, per user tenants, vectorizer modules and a built in MCP server.
Weaviate stores data objects together with one or more named vectors and searches them by meaning, by keyword (BM25) or both at once, with an alpha value that sets the balance. Vectorizer modules can create embeddings inside the database, or you supply your own vectors. Multi tenancy keeps each tenant on its own shard, and idle tenants can be set inactive or offloaded to cloud storage, which suits per user agent memory. Since the 1.37 release the server also includes an MCP endpoint at /v1/mcp with schema, tenant, hybrid search and optional upsert tools.
The latest release on October 28, 2026 was v1.40.3, a patch release with hybrid search paging fixes, and the Python client was weaviate-client 4.24.1. Code outside the repository's wl folder is BSD 3 Clause; some newer features such as Namespaces need an enterprise license key.
Pricing per weaviate.io/pricing: self hosting is free. Weaviate Cloud has a free tier (100,000 objects, 1 GB memory, 10 GB disk, one collection), a Flex plan from $45 a month billed on vector dimensions and storage, and Premium contracts from $400 a month.
Weaviate is one of 11 vector database tools on Drydock. See all vector database tools or how we check facts.
