Ultralytics Debuts Ultralytics Platform: The Definitive Way to Annotate, Train, and Deploy Vision AI

With 125,000 GitHub stars, 225 million package downloads, and 2.5 billion daily inferences, the team behind Ultralytics YOLO features a unified platform to take vision AI from raw data to production

LONDON--(BUSINESS WIRE)--Ultralytics, the company behind the YOLO family of object detection models, today introduced Ultralytics Platform, a comprehensive end-to-end vision AI platform featuring powerful SAM-powered smart annotation for building high-quality datasets in a fraction of the time; an all-new cloud training experience with 22 GPU options and real-time experiment tracking; and seamless global deployment across 43 regions with built-in production monitoring, all in a single, connected platform.

Ultralytics Platform advances every stage of the computer vision workflow. It brings together the tools developers and teams need to move from raw data to production-ready vision AI faster, with less complexity, and without switching between disconnected tools.

From open-source standard to end-to-end vision AI platform

Over the last few years, Ultralytics has built one of the most widely adopted foundations for real-time computer vision. The YOLO model family (YOLOv5, YOLOv8, YOLO11, and now YOLO26) powers more than 2.5 billion usages every day across manufacturing floors, logistics networks, healthcare systems, and autonomous vehicles. With 125,000 GitHub stars and more than 225 million Python package downloads, YOLO has become the framework that developers worldwide reach for first.

Building the models gave Ultralytics a front-row seat to a persistent problem: the models were ready for production long before the tooling around them was. Teams that could train a high-performing detector in hours were spending weeks stitching together separate services for annotation, experiment tracking, model export, and deployment, losing momentum at every handoff.

Ultralytics Platform is the answer to that gap. It extends the open-source foundation that millions of developers already rely on into a complete, connected workflow, adding managed annotation, cloud training, global deployment, and production monitoring in a single environment. Since becoming available, developers have uploaded more than 50 million images and created over 150 million annotations on the platform, a clear signal that the community that trusts Ultralytics for models is ready to trust Ultralytics for the full production workflow.

Why the model creators built the platform

Most vision AI platforms are built on third-party models. They integrate YOLO models, but they didn't build them. That distinction matters, and it delivers a fundamentally different experience.

Because Ultralytics created some of the world’s leading YOLO models, the platform is engineered from the ground up around how these models actually work, how they train, how they export, how they behave across different hardware, and where they break in production. Ultralytics YOLO26, YOLO11, and the full model family aren't add-ons or third-party integrations. They're native to the platform's architecture, from the annotation editor to the training pipeline to the deployment endpoints.

This is the core differentiator: a platform built by the same team that builds the models, with deep knowledge of every bottleneck between a labeled dataset and a live production system.

The result is a seamless workflow where annotation feeds directly into training and training feeds directly into export and deployment. Every stage is connected, with no format conversions, no tool-switching, and no infrastructure to manage.

What Ultralytics Platform delivers:

• Incredible annotation, powered by Segment Anything Model (SAM): Ultralytics Platform introduces an all-new smart annotation experience powered by the SAM. Teams can generate precise masks, bounding boxes, and oriented boxes with just a few clicks, across all five major vision tasks. Import datasets in YOLO or COCO format, upload raw images and videos, or clone community datasets to get started immediately.
• Powerful training with even more flexibility: Ultralytics Platform delivers cloud training on 22 GPU options or on local hardware with real-time metric streaming. Every experiment is automatically organized, every checkpoint preserved, and every run instantly comparable through interactive dashboards with confusion matrices, precision-recall curves, and side-by-side metrics.
• Global deployment across 43 regions: Ultralytics Platform advances deployment with dedicated, auto-scaling endpoints across 43 global regions. For teams deploying outside the platform, export to 17 validated formats, including ONNX, TensorRT, CoreML, TFLite, and OpenVINO, enables models to run natively on cloud, mobile, edge, and embedded systems. Built-in monitoring delivers full visibility into production performance.

"Most computer vision projects today never make it past the pilot stage, not because the models aren't good enough, but because the path from experiment to production is still too complex," said Glenn Jocher, Founder and CEO of Ultralytics. "We built the Ultralytics Platform to make that path simpler. One platform, from first label to live endpoint."

"We didn't set out to build another annotation tool or another training service," said Paula Derrenger, VP of Growth at Ultralytics. “We built the platform that should have existed from the beginning. It's the only end-to-end vision AI platform native to the world's most deployed object detection models. A platform designed around how vision AI actually moves from idea to production."

Pricing and availability

New users can get started with a free plan that includes signup credits and full access to annotation, training, export, and deployment. Pro plans unlock additional compute, storage, and team collaboration. Enterprise plans with custom pricing, dedicated support, and expanded capacity are available for organizations deploying vision AI at scale.

To get started with Ultralytics Platform, visit: platform.ultralytics.com

About Ultralytics

Founded by Glenn Jocher, Ultralytics is the leading force in vision AI, best known for its Ultralytics YOLO (You Only Look Once) models. With 125,000 GitHub stars, 225+ million Python package downloads, and more than 2.5 billion daily usages, Ultralytics YOLOv5, YOLOv8, YOLO11, and now YOLO26 have become widely recognized object detection models globally.

Ultralytics empowers developers and enterprises with easy-to-use, high-performance vision AI technology. Its mission is to simplify and democratize AI, making it accessible and impactful across industries, including manufacturing, agriculture, healthcare, and logistics.