What it is: MongoDB is now available as a native plugin in the marketplaces and connector directories of popular AI clients — including Claude, ChatGPT, Cursor, Devin, and Grok Build — backed by the MongoDB Atlas Managed MCP Server, a MongoDB-hosted Model Context Protocol (MCP) server. Each marketplace plugin packages the managed MCP server connection together with MongoDB agent skills, so an AI client gains both the tools to query and manage Atlas resources and the guidance to use them well, with no MCP infrastructure to host, configure, or maintain. Who it's for: This is for developers already working in AI clients such as Claude, ChatGPT, Cursor, Devin, and Grok Build who want to query, explore, and manage data in MongoDB Atlas in natural language without leaving their tool or setting up a local MCP server. It also serves Atlas administrators who need AI access to be governed, auditable, and revocable at the organization level. Why it matters: Connecting an AI client to MongoDB Atlas goes from a manual setup project — hosting a local MCP server, copying connection strings, rotating credentials — to finding MongoDB in the tool's marketplace and authorizing access with a single OAuth consent. Agent actions are attributed to the individual user who authorized it, and administrators retain control through org-level enablement and optional read-only enforcement; AI client access is disabled by default. How to get started: Click the documentation link below for a step-by-step guide on how to get started with the MongoDB Atlas Managed MCP Server in your preferred AI client.

What it is: voyage-code-4 is our next-generation code embedding model, purpose-built to improve retrieval quality and reduce costs for coding agents, supporting flexible dimensions (2048, 1024, 512, and 256) along with float32, int8, and binary quantization.Who it's for: This model is for teams building coding agents, code assistants, and code search applications over large, real-world repositories.Why it matters: Most coding agents today rely on full-text search (i.e., grep), which fails when a query describes a symptom rather than syntax; semantic retrieval with voyage-code-4 complements full-text search and significantly reduces wasted token usage, outperforming Cohere Embed v4 and Gemini Embedding 2 by 28.25% and 31.03% on agentic code retrieval at $0.12 per 1M tokens (a third below the price of voyage-code-3).How to get started: Sign up, generate a model API key, and get 200M free tokens on our latest models, then dive into the quick start to start building with voyage-code-4.

What it is: Automated embedding in Atlas Vector Search enables easily creating AI-powered Semantic Search powered by Voyage AI models. It generates, stores, and updates vector embeddings directly within MongoDB Atlas without requiring external embedding pipelines. This General Availability (GA) release introduces editable index definitions, compatibility with nested document fields, real-time token and request metrics, and automated backpressure controls under resource contention.Who it's for: This release is for developers, AI engineers, and database administrators deploying production semantic search, Retrieval Augmented Generation (RAG), and agentic workloads on MongoDB Atlas. It specifically serves teams requiring granular index management, use of different embedding models for tuning price performance, and hybrid search capabilities on complex schemas.Why it matters: The integration simplifies the developer workflow by replacing a multi-step, error-prone manual process with a single-click experience for semantic search. By handling vector synchronization and query embedding automatically, the product reduces maintenance overhead and accelerates the time to market. With the GA release, system guardrails automatically pause and resume index creation during disk pressure, ensuring operational stability and predictable resource usage during large builds.How to get started: Create or update an autoEmbed vector search index definition in the Atlas UI, MongoDB Compass, any compatible language driver, or AI Framework.