MongoDB Vector Search

Website: https://www.mongodb.com/docs/atlas/atlas-vector-search/

Description

MongoDB Vector Search extends MongoDB Atlas into a full-featured vector database. It lets you store vector embeddings alongside operational data and perform approximate or exact nearest neighbor search over those vectors. You can run semantic similarity queries, power retrieval-augmented generation (RAG) for AI applications, and combine vector search with full‑text search and structured filters in a single query. Vector Search is available on supported MongoDB Atlas clusters and integrates with popular AI frameworks and services for building intelligent, agentic systems.

Features

• Vector database capabilities

  • Store and index vector embeddings in MongoDB collections.
  • Perform approximate nearest neighbor (ANN) and exact nearest neighbor search over vectors.
  • Use vectors alongside standard BSON fields in the same documents.

• Semantic search

  • Query data by semantic similarity rather than exact keyword matching.
  • Retrieve the most relevant documents based on embedding similarity.

• Hybrid search and filtering

  • Combine vector search with MongoDB Atlas Search (full‑text search) in the same query.
  • Apply structured filters on any other document fields (e.g., metadata, categories, timestamps) together with vector similarity.

• RAG and AI use cases

  • Support retrieval‑augmented generation (RAG) by storing context embeddings and retrieving them for LLM prompts.
  • Designed to integrate into intelligent and agentic AI systems (chatbots, recommendation systems, semantic search interfaces, etc.).

• Atlas integration

  • Available on supported MongoDB Atlas clusters.
  • Uses the same operational database infrastructure as your transactional data.

• Ecosystem and tooling

  • Integrates with popular AI frameworks and services (for embeddings, LLMs, and orchestration).
  • Can be combined with other Atlas services (e.g., Atlas Search) within a unified query model.

Pricing

Pricing details for MongoDB Vector Search are not provided in the given content. It is generally consumed as part of MongoDB Atlas cluster resources; refer to the MongoDB Atlas pricing page for up‑to‑date information.