Ultralytics YOLO is the official mobile demo for running YOLO vision AI on Android.
Experience real-time, on-device computer vision with the same YOLO technology used by developers, researchers, and teams building production AI applications with Ultralytics.
Explore core YOLO tasks:
- Object detection
- Instance segmentation
- Semantic segmentation
- Image classification
- Pose estimation
- Oriented bounding boxes
Use the live camera view to switch between model tasks and sizes, adjust confidence and IoU thresholds, change camera lenses, pause inference, and share results.
The app is built with the official Ultralytics YOLO Flutter plugin and runs optimized TFLite models on Android. Nano models are included for fast first use, with larger model variants
available on demand.
Ultralytics YOLO is designed for developers, AI practitioners, students, and computer vision users who want to test YOLO capabilities directly on a mobile device.
Learn more at ultralytics.com.
Latest Version
0.7.1Uploaded by
Laudenice Vieira
Requires Android
Android 7.0+
Category
Free Tools AppContent Rating
Everyone
Security Report
Check Now
Report
Flag as inappropriateLast updated on Oct 29, 2026
Feature: Import Ultralytics Platform models by QR code and choose official or custom models in one list. Requires UltralyticsYOLO >= 8.9.17 on iOS.
Feature: Official Android models now download from the v0.7.0 release, re-exported with Ultralytics 8.4.188 (LiteRT 2.3 w8a32; QNN QAIRT 2.51, which needs qnn-runtime >= 2.51.0). Android caches refresh once.
Performance: QNN depth inference on Snapdragon 8 Elite Gen 5 drops from 26.1 ms to 7.4 ms.
Fix: Keep custom model caches distinct per export.
