We are so excited to have been featured in GitHub's Octoverse 2025 Report! 🎉 We're among the most popular open source projects for first-time contributors, alongside Ollama and ComfyUI as developers increasingly turn to AI infrastructure to experiment with models and build pipelines. As always, thank you to our amazing community for making this possible! Read more ➡️ https://ow.ly/uv6Y50Xl778
About us
Ultralytics is a leading AI company dedicated to creating transformative, open-source computer vision solutions. As creators of YOLO, the world's most popular real-time object detection framework, we empower millions globally—from individual developers to enterprise innovators—with advanced, accessible, and easy-to-use AI tools. Driven by relentless innovation and a commitment to execution, we continuously push AI boundaries, making it faster, lighter, and more accurate. Our mission is to democratize access to cutting-edge technology, providing everyone an equal opportunity to improve their lives and impact the world positively. Acta Non Verba—actions, not words.
- Website
-
http://www.ultralytics.com
External link for Ultralytics
- Industry
- Software Development
- Company size
- 11-50 employees
- Headquarters
- London
- Type
- Privately Held
- Founded
- 2022
- Specialties
- AI, Deep Learning, Data Science, Artificial Intelligence, Machine Learning, ML, SaaS, LLM, Computer Vision, and YOLO
Locations
-
Primary
Get directions
8 Devonshire Square
London, EC2M 4YJ, GB
-
Get directions
Calle de las Huertas 41
Madrid, Madrid 28014, ES
-
Get directions
9-1 Kefa Road
Shenzhen, Guangdong 518063, CN
Employees at Ultralytics
-
Ryan Leonel
Senior Full Stack Engineer | Democratizing vision AI 🚀 | ex-Bumble, ex-Meta
-
Glenn Jocher
Founder & CEO at Ultralytics | Democratizing Vision AI 🚀
-
Onuralp SEZER 👨🏻💻
Senior Machine Learning Engineer @Ultralytics, @GoogleDeveloperGroup @FedoraProject @OpenSourceExpert
-
Kirill Verbitski
Product & AI Innovation | Democratizing vision AI 🚀
Updates
-
New release v8.3.223 | FastNMS correctness 🚀 Fixes FastNMS ranking for consistent outputs, stabilizes RT-DETR validation, refactors TensorFlow/TFLite/Edge TPU/TFJS exports for reliability, and improves FLOPs accuracy for better benchmarking. Minor updates: ✅ Cleaner TensorFlow/TFLite/Edge TPU/TFJS export APIs ✅ More accurate FLOPs with `ultralytics-thop>=2.0.18` ✅ Standardized dataset class names (American English) Ultralytics v8.3.223 release notes ➡️ View the release on GitHub https://lnkd.in/daZiVmkR
-
Train Ultralytics YOLO on rare and fine-grained objects with the LVIS Dataset! LVIS (Large Vocabulary Instance Segmentation) offers 2M+ segmentation masks across 1,200+ categories, capturing real-world object frequency, from common to rare. It’s a powerful resource for: ✅ Long-tail object detection and segmentation ✅ Open-world vision and generalization ✅ Applications in robotics, retail, and autonomous systems ✅ Research in rare-object handling and instance-level understanding Supported natively in the Ultralytics ecosystem, LVIS helps you push model generalization to the next level. Explore the dataset ➡️ https://ow.ly/lzC150X01UH
-
-
New tutorial | Thresholding in image processing explained! ⚙️ Thresholding is a fundamental step in image processing, helping separate objects from their background based on pixel intensity. In this tutorial, we'll cover: ✅ The basics of thresholding and key terminology ✅ Global vs. local (adaptive) thresholding methods ✅ Real-world applications in document scanning, medical imaging, and more A simple yet powerful guide to understanding how thresholding enhances image analysis. Watch now ➡️ https://ow.ly/AOzc50XjHOB
-
-
New release v8.3.222 | TFLite INT8 for end-to-end models 🚀 Export quantized TFLite INT8 even with end-to-end models, plus more accurate quantized NMS and full IMX500 Classification support—unlocking smaller, faster deployments on mobile and edge. Minor updates: ✅ Clearer docs and hyperparameter tuning guidance ✅ IMX500 classification examples and benchmarks ✅ CI/tooling polish: stricter linting, cleaner codespell, updated Actions Ultralytics v8.3.222 release notes ➡️ Release v8.3.222 https://lnkd.in/dWbmZ-z2
-
Deploying your model is just the beginning. Here’s how to monitor and maintain it! In this guide, we outline essential best practices to ensure long-term model reliability and performance in real-world scenarios. What’s covered: ✅ Why model monitoring matters post-deployment ✅ How to identify drift, low confidence, and performance decay ✅ Tips on using tools like Prometheus and Grafana to build your own monitoring pipeline ✅ How to plan for validation, retraining, and versioning Read the full guide ➡️ https://ow.ly/QUeb50X01yf
-
-
Ultralytics reposted this
Computer Vision, Growth @ Ultralytics 🚀 | LinkedIn Instructor | Building @PixelMind | 250,000 Medium Reads | Open Source Contributor | YOLO11 | Vision language models
Community is Power, and Ultralytics Just Proved It with 100,000 Strong on LinkedIn! 🎉 The Ultralytics family has grown into a global network of 100K+ AI innovators, developers, and dreamers shaping the future of Vision AI together. This milestone is deeply personal for me. Having led this journey from the very beginning, I’m incredibly proud of how the Ultralytics brand continues to evolve, driven by collaboration, creativity, and an open-source spirit. Beyond the big milestones, countless small projects, experiments, and demos have shaped this ecosystem. Every line of code, tutorial, and community interaction has contributed to our current state. A huge thank you to everyone who has contributed, learned, shared, or believed in this mission. Your support fuels innovation every single day. Here’s to the next 100K, and to redefining what’s possible in Vision AI together! 💪 #ComputerVision #AICommunity #OpenSource #100K
-
Ultralytics reposted this
Every day, millions of developers, researchers, and innovators around the world use Ultralytics YOLO to power their ideas, from robotics and autonomous systems to manufacturing and healthcare. What unites them is a shared belief that the next breakthrough in vision AI is just a few lines of code away. YOLO Vision began as a small event in Madrid and today, it’s become a global movement: a celebration of innovation, collaboration, and community. 🌏 Grateful for everyone who made YOLO Vision Shenzhen happen, just one month after bringing YOLO Vision to London. Still smiling, thinking about how far Ultralytics, and especially the YOLO Vision idea, has come since 2022. None of this would have been possible without our incredible partners and sponsors who helped bring this vision to life: Hailo, Infermove, D-Robotics, Moore Threads Technology Co. Ltd., Baidu, Inc., RWVK, Model Scope, Seeed Studio, DEEPX, M5Stack, SegmentFault 思否, and Extreme Vision 极视角科技; along with all the companies, users, and researchers who joined us to celebrate the future of vision AI. A heartfelt thanks to Zinnia Louise Pourdad, Ziyue Wang, and Jayln, who made everything run seamlessly, and to the entire Ultralytics team working tirelessly behind the scenes. And a special nod to Cinco Jotas from Grupo Osborne, who brought a taste of Spain to China, a small but meaningful reminder of Glenn Jocher and my Spanish roots. ✨ And for everyone following our journey, stay tuned to our socials and GitHub for the official launch of the Ultralytics Platform and YOLO26! Glenn Jocher
-
-
-
-
-
+15
-
-
Ultralytics reposted this
🌟Wrap up for YOLO Vision Shenzhen 2025!!🌟 We're beyond excited to have been part of #YOLO #Vision Shenzhen 2025 – what a journey it's been! 🎉 💫 The AI Sensing team had a fantastic time at the event, diving deep into meaningful conversations and staying up-to-date with the latest trends and innovations in the industry. 📸 We proudly introduced our latest product, the #reCamera PoE, through an interactive #workshop. This hands-on session allowed us to engage directly with users, answer questions, and demonstrate the power of our product in real-time. 🛠️ It was amazing to see how participants quickly grasped the capabilities of the #reCameraPoE and used it to create innovative, functional projects on the spot! 🌐 We left with a wealth of new knowledge, fresh ideas, and a stronger connection to the community. Seeed can't wait to continue this journey of growth and collaboration! cc: Ultralytics #SeeedStudio #Ultralytics #AI #AIoT #VisionAI #YOLOVision #Hailo #DEEPX #HowtomakeanAICamera
-
-
YOLO Vision Shenzhen 2025: That’s a wrap! ⭐ It’s the first year we bring our annual event to China and it has been incredible fun! From keynotes and demos to networking and holding live workshops, there is nothing to mark progress like seeing our team and community grow together! Thank you to our sponsors and team for having made this milestone possible! See you next year! Rewatch now ➡️ https://lnkd.in/ddGJVzmP DEEPX Baidu, Inc. D-Robotics Intel Corporation Seeed Studio RWKV Project (under LF AI & Data Foundation) SegmentFault 思否 摩尔线程智能科技（北京）有限责任公司 M5Stack Datawhale Qwen Infermove Hailo
-