MongoDB Atlas Vector Search
Free tierBuild intelligent applications powered by semantic search and generative AI with a native, full-featured vector database.
Key strengths
Facts verified against mongodb.com on
Recent changes
- Plans updated: Free tier at $0/hour (512MB storage, shared RAM/vCPU). Flex tier at $0.011/hour (up to $30/month, 5GB storage). Dedicated tier starting at $0.08/hour ($56.94/month, 10GB–4TB storage). Atlas Infinite at $0.09/hour (public preview).
- Models updated: Voyage AI
- Self-hosting: now available
Get alerts when MongoDB Atlas Vector Search changes
Free. We'll email you when pricing, the free tier, API access or open-source status changes.
MongoDB Atlas Vector Search — Technical Overview
MongoDB Atlas Vector Search enables developers to build AI-powered applications directly on their live operational data in MongoDB — eliminating the need for a separate, dedicated vector database. It supports vector search, hybrid search, and generative AI workloads (including RAG pipelines and agentic systems) within the same MongoDB cluster that already stores your application data.
Key architectural strengths:
- Unified data layer: Vector embeddings live alongside operational documents — no ETL pipelines or data synchronization required.
- Automated Embeddings: Atlas manages embedding generation via integrated Voyage AI models, removing the need to orchestrate external embedding services.
- Hybrid search: Combine vector similarity search with lexical (full-text), geospatial, and MongoDB aggregation pipeline queries in a single request.
- Independent scaling: Dedicated Search Nodes allow vector search compute to scale independently from your operational database workload.
- Native reranking: Built-in support for Voyage AI reranker models to improve result relevance post-retrieval.
- Deployment flexibility: Available as a fully managed cloud service (Atlas) or self-hosted via Enterprise Advanced and Community Edition.
- Integrations: Voyage AI (embeddings & reranking), Okta (identity & access management).
Pricing tiers:
| Tier | Rate | Notes |
|---|---|---|
| Free | $0/hour | 512 MB storage, shared RAM/vCPU |
| Flex | $0.011/hour | Up to $30/month, 5 GB storage |
| Dedicated | from $0.08/hour | $56.94/month, 10 GB–4 TB storage |
| Atlas Infinite | $0.09/hour | Public preview |