MongoDB Acquires Voyage AI To Add Embedding Models
Written by Kay Ewbank
Monday, 10 March 2025
MongoDB is to acquire Voyage AI with the intention of using Voyage AI's facilities within MongoDB so developers can build apps that use AI more reliably. Voyage AI produces embedding and reranking models.
MongoDB is a NoSQL document database that stores its documents in a JSON-like format with schema. MongoDB Atlas is the fully-managed cloud database from the MongoDB team.
The Voyage AI range provides intelligent retrieval models that underpin retrieval-augmented generation (RAG) and reliable LLM applications, according to the company.
Explaining the acquisition, Dev Ittycheria, President and CEO, MongoDB, said that embedding generation and reranking are two key AI components that capture the semantic meaning of data and assess the relevance of queries and results.
"We believe embedding generation and re-ranking, as well as AI-powered search, belong in the database layer, simplifying the stack and creating a more reliable foundation for AI applications."
With Voyage AI, MongoDB solves this challenge by making AI-powered search and retrieval native to the database. MongoDB says Voyage AI has built a world-class AI research team with roots at Stanford, MIT, UC Berkeley, and Princeton and has become a leader in high-precision AI retrieval.
The integration with MongoDB will happen in three phases. In the first phase, Voyage AI's text embedding, multi-modal embedding, and reranking models will remain widely available through Voyage AI's current APIs and via the AWS and Azure Marketplaces.
Voyage AI's capabilities will then be embedded into MongoDB Atlas, starting with an auto-embedding service for Vector Search, which will handle embedding generation automatically. Native re-ranking will follow, and MongoDB also plans to expand domain-specific AI capabilities for different industries such as financial services and legal, and for specific uses such as code generation.
The final move will be to add enhanced multi-modal capabilities for retrieval and ranking of text, images, and video. MongoDB also plans to introduce instruction-tuned models so developers will be able to refine search behavior using simple prompts.
Microsoft has released Hyperlight Wasm, a Hyperlight virtual machine (VM) "micro-guest" that can run Wasm component workloads written in many programming languages.
Andrew Barto, Professor Emeritus of Information and Computer Sciences at the University of Massachusetts, Amherst and Richard Sutton, a professor of Computer Science at the University [ ... ]
