ultralytics/yolo-ios-app

Ultralytics YOLO iOS app and Swift package for real-time Core ML inference across major computer vision tasks.

What it solves

This project provides a way to run high-performance YOLO vision models on iOS and iPadOS devices. It enables real-time, on-device inference for various computer vision tasks without requiring a constant internet connection or external servers.

How it works

The project is built using Swift and Apple's Core ML framework, allowing it to run directly on the Apple Neural Engine (ANE) and GPU. It supports YOLO11 and YOLO26 models, with the latter being NMS-free to optimize performance. The project is delivered as both a complete demonstration app and a lightweight Swift package that developers can integrate into their own apps via Swift Package Manager or CocoaPods.

Who it’s for

  • iOS Developers: Those looking to add real-time object detection or segmentation to their apps.
  • AI Researchers/Users: People who want to test YOLO models on mobile hardware using a native app.

Highlights

  • Multi-task Support: Handles object detection, instance and semantic segmentation, depth estimation, classification, pose estimation, and oriented bounding box (OBB) detection.
  • High Performance: Achieves camera-rate (~30 FPS) real-time inference on recent iPhones.
  • Zero Dependencies: Built using pure Swift and Apple's first-party frameworks.
  • Flexible Integration: Offers a ready-to-use app, a Swift library, and example applications for developers.

Related

  • Project
    john-rocky/CoreML-Models

    A comprehensive zoo of pre-converted CoreML models for on-device AI on iOS and Android, covering everything from computer vision to LLMs.

  • Project
    FaceAISDK/FaceAISDK_iOS

    An on-device, fully offline SDK for iOS that provides face detection, recognition, and liveness detection to enable secure identity verification without network access.

  • Project
    ml-explore/mlx-swift-lm

    A Swift package for building applications with large language models and vision language models using MLX Swift, featuring model loading, fine-tuning, and guided generation.

  • Project
    ultralytics/ultralytics

    A comprehensive computer vision framework featuring YOLO models for real-time object detection, segmentation, classification, and pose estimation.

  • Project
    Geekgineer/YOLOs-CPP

    A production-ready C++ inference engine that provides a unified API for the entire YOLO family, supporting tasks from object detection to metric depth estimation.