MongoDB, announced a significant expansion of its artificial intelligence capabilities at MongoDB.local San Francisco, introducing an industry-first approach that unifies its core database platform with Voyage AI’s industry-leading embedding and reranking models. This integration forms a unified data intelligence layer for production AI. It helps developers build, deploy, and scale advanced AI applications. This way, they can reduce hallucinations and avoid moving or duplicating data across systems.
To help organizations confidently take AI applications from experimentation into production, MongoDB unveiled a comprehensive set of new AI capabilities designed to streamline development and operations. These include five new embedding models from Voyage AI, MongoDB’s expanded embedding and retrieval model suite, Automated Embedding for MongoDB Community Vector Search, new embedding and reranking AI model APIs within MongoDB Atlas, and an AI-powered data operations assistant for MongoDB Compass and Atlas Data Explorer. Together, these innovations reinforce MongoDB’s position as a leading AI-ready data platform, trusted by more than 60,000 customers worldwide to run mission-critical workloads. Voyage AI models are accessible through MongoDB Atlas via API, integrated with MongoDB Community through managed Automated Embedding, and continue to be available as a standalone offering independent of MongoDB.
“The biggest challenge customers face with AI isn’t experimentation, it’s operating reliably at scale,” said Fred Roma, Senior Vice President of Product and Engineering at MongoDB. “Developers want fewer moving parts and clearer paths from prototype to production. With today’s launches, MongoDB is raising the bar, helping teams reduce complexity and focus on building AI applications that perform in real-world, mission-critical environments.”
Also Read: AWS IoT Device Management Adds Wi-Fi Simple Setup
Transforming data into AI intelligence
As AI initiatives transition into production, many organizations are discovering that traditional data architectures were never designed to support context-aware, retrieval-intensive workloads at scale. Developers are often forced to manage fragmented stacks that combine operational databases, vector stores, and external model APIs—introducing unnecessary complexity, increased latency, and heightened operational risk at a time when reliability and speed are critical. This fragmentation has emerged as a major obstacle to AI innovation, directly impacting customer-facing applications and business outcomes.
MongoDB addresses these challenges by consolidating the essential capabilities required to build and operate AI applications in production within a single data platform. Rather than stitching together disparate systems, teams can keep operational data, vector search, and AI model integrations tightly aligned. This unified approach reduces synchronization overhead, lowers latency, and simplifies architecture—enabling faster iteration and AI applications that are designed for reliable, real-world production use, not just proofs of concept.
Key new capabilities include:
State-of-the-art accuracy with models from Voyage AI: The general availability of the new Voyage 4 series delivers high-performance embedding models that outperform Gemini and Cohere on the public RTEB leaderboard, offering more accurate retrieval at lower cost. The Voyage 4 lineup includes the balanced, general-purpose voyage-4 model; the flagship voyage-4-large model for maximum retrieval accuracy; voyage-4-lite for optimized latency and cost efficiency; and the open-weights voyage-4-nano model designed for local development, testing, and on-device use.
Enhanced context extraction from video, images, and text: The general availability of the voyage-multimodal-3.5 model expands multimodal support beyond interleaved text and images to now include video. Building on Voyage AI’s earlier production-grade multimodal models, voyage-multimodal-3.5 more effectively vectorizes complex data types—such as tables, graphics, figures, slides, PDFs, and videos—into a unified semantic representation. This significantly reduces the need for complex document parsing workflows, improving retrieval accuracy and helping developers deliver more reliable, context-aware AI applications.
Automated Embedding for MongoDB Vector Search: With Automated Embedding, MongoDB can automatically generate and store high-fidelity embeddings using Voyage AI whenever data is inserted, updated, or queried. By handling embedding generation natively within the database, MongoDB eliminates the need for separate embedding pipelines or external model services. Embeddings stay continuously up to date as data changes, ensuring accurate retrieval and consistent context for AI-driven applications. Automated Embedding is currently available in public preview, with support across MongoDB drivers—including JavaScript, Python, and Java—and popular AI frameworks such as LangChain and LangGraph (Python). It is available today for MongoDB Community, with MongoDB Atlas support coming soon.
“We were looking for extremely accurate embedding models, and Voyage AI provided accuracy at scale,” says Sudheesh Nair, Cofounder and CEO of TinyFish. “The Python APIs that Voyage comes out of the box with are also extremely lightweight and very fast.”
“Today, companies need to move extremely fast, and at very lean startups, you need to only focus on what you are building,” said Rotem Weiss, CEO of Tavily. “MongoDB allows us to focus on what matters most, our customers and our business.”