MongoDB adds MCP server, expands AI development capabilities
While the database vendor isn't breaking new ground with its updated features, they combine to provide customers with valuable tools that will help to build modern applications.
MongoDB unveiled new capabilities aimed at better enabling customers to develop AI tools such as the public preview of vector search in its self-managed offerings and the general availability of its Modern Context Protocol server.
In addition, the vendor launched an updated version of its document database and an AI-powered service that helps customers modernize their data infrastructure to meet current enterprise demands.
The new capabilities were revealed on Tuesday and Wednesday during MongoDB.local NYC 2025, a user conference in New York City.
While not innovative features on an individual basis, collectively MongoDB's new capabilities are valuable additions for its customers, according to Kevin Petrie, an analyst at BARC U.S.
"MongoDB's product direction is right on target," he said. "MongoDB continues to help companies modernize and build applications that drive smart, multifaceted workflows. While the individual announcements are incremental, together they add up to a significant step forward for companies that need to modernize legacy apps, boost performance and feed rich inputs to AI applications."
Stephen Catanzano, an analyst at Enterprise Strategy Group, now part of Omdia, similarly noted the strategic importance of the new capabilities as MongoDB transitions from database specialist to broader data platform vendor.
"These announcements represent a strong, strategic set of updates that enhance MongoDB's value proposition, particularly in AI integration and application modernization," he said. "The combination … helps customers accomplish more with their MongoDB deployments beyond mere incremental improvements. Like many, they are moving toward being a platform player and less of a database."
Based in New York City, MongoDB provides a NoSQL database platform designed to handle the rising scale of enterprise data and AI workloads that traditional relational databases sometimes struggle to manage.
New capabilities
Enterprises have been increasing their investments in AI development ever since OpenAI's launch of ChatGPT marked a significant improvement in generative AI (GenAI) technology.
GenAI has the potential to make workers better informed and more efficient in several ways. It enables nontechnical employees to analyze data using natural language chatbots and relieves trained experts of certain work, such as code generation. Now, AI development is evolving beyond GenAI to include agents. Agents are applications with reasoning capabilities that enable them to act autonomously to further assist workers by surfacing insights on their own and taking on more complex tasks.
Because data is the "intelligence" in AI, many data management vendors have responded to the AI development surge by adding tools that simplify training AI models and applications using proprietary data.
Now, MongoDB is adding text search and vector search capabilities that are key to the AI development process to its on-premises Enterprise Server and self-managed Community Edition platforms. Previously, such capabilities were available on only MongoDB's cloud-based Atlas platform.
"The extension of search and vector search capabilities to Community Edition and Enterprise Server stands out as particularly significant because it democratizes AI development across all deployment environments without requiring Atlas," Catanzano said.
Petrie likewise noted the importance of extending such capabilities to all MongoDB customers, given that BARC research shows that about one-third of AI workloads run on-premises. Apprehension about data security in the cloud, data migration costs and data sovereignty concerns are among the reasons enterprises choose to keep certain workloads out of the cloud, according to Petrie.
This meets a critical requirement for companies that need higher certainty levels than typical semantic search can provide, as it retrieves unstructured content for GenAI workflows, especially retrieval-augmented generation.
Kevin PetrieAnalyst, BARC U.S.
"I like how MongoDB is combining keyword and vector search for on-premises environments," he said. "This meets a critical requirement for companies that need higher certainty levels than typical semantic search can provide, as it retrieves unstructured content for GenAI workflows, especially retrieval-augmented generation."
Meanwhile, the impetus for extending text and vector search to on-premises environments came partially from conversations with customers, according to Ben Cefalo, MongoDB's senior vice president and head of core products, who spoke during a virtual press conference Monday.
"Definitely customer feedback," he said. "Also, we truly believe in the 'run anywhere' moniker, so while we built Atlas Search and Atlas Vector Search, that was missing from our on-premises offerings."
Beyond making search capabilities more widely available, MongoDB is joining the parade of vendors adding support for Model Context Protocol (MCP) by making MongoDB MCP Server generally available.
Launched by Anthropic in November, MCP is an open framework that simplifies agent development by standardizing how agents interact with the data sources used to train them, including LLMs. AWS, Google Cloud, Microsoft, Oracle, Databricks, Snowflake and Informatica are among the many other data management vendors now providing MCP support.
Underlying the added AI development capabilities is MongoDB 8.2, the latest version of the vendor's database platform. MongoDB released version 8.0 of its platform in October 2024. The update improves performance to meet the demands of AI workloads, including faster unindexed queries, in-memory reads and higher throughput.
"MongoDB is positioning itself effectively against traditional database competitors and specialized startups by offering a unified platform that handles traditional operational workloads alongside AI capabilities," Catanzano said.
While enhanced search capabilities, MCP support and improved performance aid customers with AI-ready infrastructures, many enterprises have outdated technology frameworks that are incapable of meeting modern data and AI workload demands.
According to a 2022 report by the Consortium for Information and Software Quality, poor software quality cost U.S. enterprises a combined $4 trillion. Software failures, lost developer time and systems maintenance were key factors affecting software quality.
MongoDB's Application Modernization Platform (AMP), developed over two years in conjunction with certain MongoDB customers and now generally available, helps customers modernize with a combination of AI-powered software tools, a set of best practices and dedicated AMP engineers.
Regarding why MongoDB developed its AMP, exploding interest in AI and the resulting need for more advanced infrastructures played a role, according to Shilpa Kolhar, the vendor's senior vice president of product and engineering.
"It's a timing thing," she said during the press conference. "In the last couple of years, we have helped customers in various industries migrate and modernize from their legacy tools to modern applications built on MongoDB. We feel that now is the time we can help our customers wherever they are to get them to be ready to embrace new AI technologies."
Looking ahead
While MongoDB's new capabilities address customer needs and help keep the vendor's platform competitive, there is still more MongoDB can do to serve the needs of its users and perhaps even attract new ones, according to the analysts.
Petrie noted that many enterprises now turning some of their operations over to agents are concerned about surrendering control. Therefore, improved AI governance is one way MongoDB could better aid its users, he suggested.
"AI adopters are rightfully concerned about the risks of agentic AI, and uncertain how to mitigate those risks," Petrie said. "MongoDB should strengthen its governance capabilities. For example, its application metadata can support modern governance programs, helping enforce policies and controls to data, model, and agent activities."
Catanzano, meanwhile, recommended that MongoDB provide users with industry-specific templates to better enable customers in verticals such as financial services, retail and healthcare to more easily develop AI tools.
"Creating prebuilt components and reference architectures would lower the barrier to entry for organizations looking to implement AI solutions while showcasing MongoDB's versatility beyond traditional database workloads," he said.
Eric Avidon is a senior news writer for Informa TechTarget and a journalist with more than 25 years of experience. He covers analytics and data management.
