MongoDB adds AI retrieval tools to Atlas platform

MongoDB has introduced new AI data retrieval features to its Atlas platform, including integrations with popular developer tools such as ChatGPT, Claude, Codex, Grok Build, and Devin. The update incorporates Voyage AI models directly into the database and introduces a managed Model Context Protocol (MCP) server, which allows coding agents to connect to live operational data using existing access controls.

By integrating retrieval models and direct links to coding assistants within Atlas, MongoDB aims to position its operational database as the essential memory and retrieval layer for AI applications. This approach addresses the common issue of data synchronization errors that occur when organizations move information from operational databases to separate vector stores. With these updates, developers can query and modify data using natural language or through automated processes without managing extra infrastructure.

Key technical additions include Automated Embeddings in Atlas, an Embedding and Reranking API, and the voyage-code-4 model specifically designed for software repository retrieval. These tools enable real-time context for AI agents, ensuring they act on live application data rather than stale information, thereby simplifying the AI infrastructure stack for developers.

(Source：It Brief New Zealand)