MongoDB launched Atlas Agent Engine, a unified execution, memory, and governance layer for production AI agents. While AI agents prove their value quickly in proof-of-concept testing, productionizing them remains a major bottleneck. Doing so requires accurate retrieval, persistent memory, and enterprise-grade security and governance. Without a single platform, engineering teams must stitch together disparate tools that break every time underlying models or frameworks evolve. That’s what Atlas Agent Engine is designed to solve for.
Atlas Agent Engine gives teams a modular way to put agents into production. Retrieval is powered by MongoDB Voyage AI, whose embedding and reranking models rank among the top performers on RTEB, a benchmark built to reflect real enterprise retrieval instead of academic datasets. Customers can adopt the memory and governance layers independently or with the runtime, using existing models and frameworks they know. Atlas Agent Engine is available in public preview, with consumption-based pricing for Atlas Agent Runtime and Atlas Agent Memory. Usage draws on customers’ existing Atlas commitments, so adoption extends infrastructure already in place rather than requiring a new contract.
Built to get agents into production
Enterprises building agents often run into the same three challenges: actions nobody can govern, agents that forget, and lock-in to a single model or framework. Atlas Agent Engine solves all three, grounded in the same operational platform that more than 70,000 customers already run on, with enterprises like Paysafe already building toward production.
“Organizations that want to put agents in production are being forced into a false tradeoff: either adopt one vendor’s runtime and accept being locked into a model and cloud, or piece together a framework and manage governance and memory on their own,” said Pablo Stern-Plaza, Chief Product Officer, AI and Emerging Products, MongoDB. “With the launch of Atlas Agent Engine, that false tradeoff ends today. Enterprises get the real-time context their agents need, with governance and security built in from the start, and the freedom to run any model, any framework, and on any cloud. We didn’t want to ask customers to predict the future. We wanted to build something that works no matter what they choose.”
Governed by default. Most platforms handle identity, audit, guardrails, and cost controls as separate systems teams have to stitch together themselves. Atlas Agent Engine puts it all behind one control plane: every action is logged against a real identity, human or agent, and governed by policy that can’t be quietly switched off. Governance is built in, not bolted on after launch. So when someone asks what an agent did and who authorized it, the answer takes seconds, not weeks. And because governance, memory, and retrieval run as one system instead of stitched-together services, there’s less to secure and fewer places for things to break.
Memory and retrieval built in. Without built-in memory, agents start every conversation from zero, and teams end up rebuilding memory infrastructure for every new agent. Atlas Agent Engine builds memory into the platform itself, using Voyage AI embeddings and MongoDB’s native retrieval, so agents get more accurate while spending fewer tokens.
Open design. Standardizing on one model, cloud, or framework is one of the riskiest infrastructure bets a leader can make in a market that moves this fast. Atlas Agent Engine is neutral across AI models and frameworks. Because it’s built on open standards like MCP and A2A, changing course later takes a configuration change rather than an expensive rebuild. It will also run across any cloud, self-managed or even a laptop, so the same agent works everywhere without rebuilding cloud by cloud. Atlas Agent Engine adds governed execution, memory, and cost control on top of what teams already run, rather than asking them to replace it.
MongoDB is committed to giving customers the openness and security they need. We are joining the Linux Foundation’s Open Secure AI Alliance and Agentic AI Foundation to drive open software and standards for secure, interoperable agents—so organizations can move agents into production with the flexibility and control to scale as their needs evolve.
The company also announced MongoDB 9.0, the best version of MongoDB ever built, and Atlas Infinite, the largest architectural innovation to MongoDB Atlas since its inception in 2016. MongoDB 9.0 is the next-generation engine underpinning MongoDB Atlas, the company’s fully managed cloud data platform, Enterprise Advanced, and Community Edition, delivering superior performance wherever a customer’s data lives. Atlas Infinite extends that foundation with the elasticity modern applications and AI agents demand, making MongoDB the intelligent data platform for the AI era.
