Weaviate is an open-source database for AI apps that keeps data objects together with their embeddings, and supports meaning-based, keyword and hybrid search.
In plain words
Weaviate keeps each record, such as an article, next to its vector embedding. It can create those embeddings for you when you import data, using built-in connections to model providers such as OpenAI, Cohere, Hugging Face and Google. Searches can use semantic similarity, traditional BM25 keyword matching, or both, plus structured filters. It also has built-in RAG and reranking. It is written in Go and has client libraries for several languages.
A quick example
Weaviate's own starter example inserts three short sentences, about vector databases, embeddings and hybrid search, into a collection and then runs a semantic search over them. Results are ranked by closeness in meaning to the query rather than by shared words alone.
Where you'll meet it
- Semantic and hybrid search.
- RAG applications with built-in generation.
- Image search.
Related
Sources
- Docsweaviate/weaviate, Weaviate (GitHub) · read 29 Sept 2026
- DocsWeaviate Database, Weaviate documentation · read 29 Sept 2026
Checked by WhatIsAI against the sources above · 29 Sept 2026. Spotted an error? Report it.