MongoDB Unveils Voyage AI and Expands AI Partnership Ecosystem

MongoDB today announced a major leap forward for AI-powered development at the Ai4 2025 conference, unveiling Voyage AI models and bolstering its partner ecosystem to streamline AI adoption. These innovations aim to deliver faster, more accurate, and cost-effective AI across enterprise applications.

Voyage AI introduces cutting-edge embedding models designed to deliver context-aware and highly relevant data retrieval—eliminating the need for complicated metadata pipelines or chunking workarounds.

Standouts include:

• voyage‑context‑3: Captures full document context to enhance relevancy and reduce sensitivity to text partitioning.
• voyage‑3.5 & voyage‑3.5‑lite: Offer top-tier retrieval accuracy while maintaining excellent price-performance.
• rerank‑2.5 & rerank‑2.5‑lite: Feature instruction-driven re-ranking to boost retrieval accuracy across benchmarks.

MongoDB also introduced the MongoDB MCP Server, now in public preview. This protocol simplifies the connection between MongoDB and AI tools such as GitHub Copilot, Anthropic Claude, Cursor, and Windsurf—enabling natural language querying and streamlining data-driven development.

Building on the improved data and tool access, MongoDB has expanded its AI partner ecosystem with key collaborators:

• Galileo: Adds reliability and observability for AI agents.
• Temporal: Enables durable execution for resilient, complex workflows.
• LangChain: Enhances support for GraphRAG and natural language querying, facilitating agentic AI development.

MongoDB highlighted strong marketplace momentum, citing that over 8,000 startups and leading enterprises are leveraging its AI tools. Additionally, the MongoDB Atlas platform now attracts more than 200,000 new developer registrations each month, reflecting growing demand.

"Many organisations struggle to scale AI because the models themselves aren't up to the task. They lack the accuracy needed to delight customers, are often complex to fine-tune and integrate, and become too expensive at scale," said Fred Roma, SVP of Engineering at MongoDB. "The quality of your embedding and reranking models is often the difference between a promising prototype and an AI application that delivers meaningful results in production. That's why we've focused on building models that perform better, cost less, and are easier to use—so developers can bring their AI applications into the real world and scale adoption."