MongoDB has announced a set of capabilities at its MongoDB.local London 2026 event that aim to simplify the data infrastructure required for running AI agents in production. The updates include automated embeddings generation, persistent agent memory, a new database version with performance improvements, and expanded deployment options.
Overview
The core problem MongoDB is addressing is the complexity of stitching together separate systems for vector search, memory, and real-time operational data. The company now offers a single platform that combines a real-time database, full-text and vector search, memory, embeddings, and reranker models. This eliminates the need for teams to maintain external pipelines and synchronize data across multiple tools.
What's new
Automated Voyage AI Embeddings in MongoDB Vector Search (public preview) – Embeddings are now generated automatically as data is written or updated. This removes the manual infrastructure work of building and maintaining search infrastructure. MongoDB states that its Voyage AI embedding models rank first on the Retrieval Embedding Benchmark (RTEB).
LangGraph.js Long-Term Memory Store (generally available) – JavaScript and TypeScript developers now have access to persistent, cross-conversation agent memory, powered by MongoDB Atlas as a single backend. No additional database is required.
MongoDB 8.3 (generally available) – Delivers up to 45% more reads, 35% more writes, 15% more ACID transactions, and 30% more complex operations compared to MongoDB 8.0, without requiring application code changes. Common data transformations have been moved into the database itself, reducing the need for external pipelines.
Cross-region connectivity for AWS PrivateLink (generally available) – Database traffic between MongoDB Atlas clusters in different AWS regions stays on the AWS private network, with no exposure to the public internet.
Feast Feature Store Integration with MongoDB (generally available) – Allows teams to use Feast for feature management while storing feature data in MongoDB.
New Query Expressions for Data Transformation (generally available) – Enables data transformations directly within the database.
Deployment flexibility
MongoDB runs across AWS, Google Cloud, Microsoft Azure, on-premises, and in hybrid environments. Customers get one database, one API, and one set of skills that work consistently across deployment targets. This is particularly relevant for banks, healthcare organizations, and government agencies with data residency requirements.
Bottom line
MongoDB's latest updates reduce the operational overhead of building AI agents by consolidating vector search, memory, and real-time data into a single platform. For teams currently managing separate systems for embeddings, memory, and search, the new capabilities could cut weeks of infrastructure work. The performance improvements in MongoDB 8.3 apply to existing workloads without code changes, making it a straightforward upgrade for current users.