AI is moving from generating answers to taking action. Your data platform should move with it. At #MongoDBlocal NYC, we announced the biggest expansion of the MongoDB platform yet. New capabilities help you: 🚀 Run on MongoDB 9.0, now GA with up to 2x higher throughput ♾️ Scale compute and storage independently with Atlas Infinite 🤖 Build, deploy, and govern agents on live data with Atlas Agent Engine ⏱️ Let agents spin up databases in seconds with Ephemeral Clusters 🔐 Run prefix, suffix, and substring queries on fully encrypted data 🎯 Rerank results in the aggregation pipeline with Voyage AI rerank-3 Everything you need to take AI apps and agents to production, on one intelligent data platform. mongodb.social/6019BGt1h7

📢 Introducing rerank-3 and rerank-3-lite, the next generation of Voyage rerankers. A drop-in upgrade to the Rerank 2.5 series models, with focused improvements on long document and code retrieval. 📈 On standard retrieval, rerank-3 outperforms Cohere Rerank v4.0 Pro by 2.72% and Qwen3-Reranker-8B by 3.02%. 📚 On long document retrieval, rerank-3 outperforms rerank-2.5 by 3.35% and Cohere Rerank v4.0 Pro by 13.84%. 💰 rerank-3-lite now matches the retrieval quality of rerank-2.5 at 40% of the price. Try it today. The first 200M tokens are free. mongodb.social/6016BGsoXp

Congratulations to the winners of The Harness Engineering & Model Wrangling Hackathon at #MongoDBlocal NYC! 🎉 🥇 Ripple: designs circuit boards from plain English, grades them against hidden checks, and evolves its own harness from the failures 🥈 Offload: learns the work you repeat, distills it into tested skills, then evolves its own harness against a held-out task suite 🥉 MorgaDarsha: local-first desktop automation harness that verifies outcomes locally and shares what it learns across agents through a MongoDB memory layer Thank you to everyone who built and demoed in New York, to our judges, partners, and to @cerebral_valley for helping make the hackathon possible.

Builders need an ecosystem of tools, models, frameworks, and services that can move AI from prompt to production in the real world, faster. MongoDB is proud to provide the technical foundation for that work. The AI Builder ecosystem plays a critical role in delivering this value. This week at #MongoDBlocal NYC we announced new partnerships that enable this new era of AI application development with Emergent, Monk.io, and Retool. Together we are enabling the next generation of applications and builders! mongodb.com/company/blog/eve…

POV: you spent the day at #MongoDBlocal NYC with 1,200+ builders and somehow ended up with more pins than you can count. Some of our biggest announcements yet hit the main stage, customers and partners showed us what they're building with MongoDB, and our Hackathon Champions took home their trophies after going head-to-head with 250+ participants. Here's what Pier 36 looked like from the floor 👇

“Atlas Agent Engine brings the enterprise-ready capabilities, context and constraints needed to help AI agents deliver real-world impact. Combined with Accenture’s governance, architecture, and deep industry expertise, it creates a powerful foundation for accelerating AI transformation and delivering outcomes at scale. Our shared commitment to delivery and customer success makes this partnership particularly strong.” - Ram Ramalingam , Global Lead, SW Engineering & Head of RDE @Accenture

Before we sell agentic AI to anyone else, we put it to work on ourselves. Holly is MongoDB's internal AI assistant, built on Atlas Agent Engine, designed to connect the tools our sales team already uses into one workspace, instead of ten. What it's done for research time, deal prep, and account strategy might surprise you. ✨ Read the full use case here: mongodb.social/6014BGtGIn

Live from the floor at #MongoDBlocal NYC Sessions with @CDKGlobal , @okta, @coinbase, and @FireworksAI_HQ just delivered. Real stories, real scale, and plenty of insights on building with MongoDB.

Building an AI agent demo is one thing. Building a system your business can depend on takes more. “You can’t just throw $100,000 at a tough programming challenge and actually have it build production software.” — @walden_yan, Co-Founder & CPO, @cognition At #MongoDBlocal NYC, leaders from @cognition, @temporalio, and @openrouter joined @sophiadew for “Building at the Frontier”—a candid conversation about what it takes to run AI agents in production, from testing and verification to recovering from failures and controlling access to data. The takeaway: As agents gain autonomy, the data layer matters more than ever. Watch the full session: mongodb.social/6013BGsFdF