Vector Search in MongoDB (www.mongodb.com)

MongoDB announced a public preview of native search and vector search for self-managed deployments—now available in MongoDB Community Edition and Enterprise Server—bringing full-text and semantic/vector search directly into the core database. The release exposes $search, $searchMeta and $vectorSearch aggregation stages with functional parity to Atlas (excluding preview-only features), enabling developers to build relevance-based features, semantic search, RAG pipelines, hybrid keyword/vector queries, recommendations and agentic workflows without bolting on separate search or vector DBs. That consolidation aims to cut architectural complexity, operational overhead, and consistency problems caused by managing and syncing multiple systems. Technically, the Community preview targets MongoDB 8.2+ and runs via a separate mongot binary (also available as a container), while Enterprise search nodes are deployed self-managed in Kubernetes and connect to existing clusters (operator/version requirements noted for 8.0.10+/8.2+). The features integrate with frameworks like LangChain, LangGraph and LlamaIndex to streamline RAG and LLM app development. The public preview is free for testing; Community capabilities will be included under the SSPL at no extra cost, and Enterprise will join a paid subscription later. MongoDB plans iterative updates based on feedback during the preview.