JumpSense AI: The Ultimate On-Device Jump Rope Counter and Form Analyzer

Overview

JumpSense AI is a specialized fitness utility designed to transform an iPhone into an AI-powered jump rope coach. By leveraging Apple's built-in Vision framework, the app provides real-time jump counting, double-under detection, and form analysis entirely on-device. It requires no wearables, no manual tapping, and no recurring subscriptions.

Benefits

JumpSense AI offers several key advantages for anyone looking to improve their jump rope skills. The app counts every jump instantly by detecting body pose changes frame-by-frame. This means users get live updates without any delays or the need for manual input. Unlike many fitness apps that rely on cloud servers, JumpSense AI processes all data locally on the iPhone. This ensures zero network access is needed, which means the app works entirely offline. It also guarantees that no personal data is collected, tracked, or uploaded to the cloud. The processing happens in real-time without any lag.

The app can automatically identify double-unders by analyzing the time interval between jumps. Users can tune the interval threshold to match their specific speed. This allows the app to separate single-unders from double-unders in the statistics. A real-time display shows jumps per minute, helping users pace intervals, hit target speeds, and measure fitness progress. JumpSense AI offers three distinct display modes to suit different training needs. Count Mode features a large jump count HUD with a skeleton overlay. Rhythm Mode displays the count alongside a live signal waveform that visualizes the rhythm of each jump. Form Mode shows a skeleton-only view, which is ideal for detailed posture and form analysis without distraction.

The app includes built-in hip-baseline tracking that filters out camera movement. This ensures accurate counting even if the phone is handheld or placed on a shaky stand. Users can record their jump sessions with the skeleton overlay and count HUD baked directly into the video. The overlay is perfectly synced to the camera frame with no lag. Recordings can be saved directly to the Photos library for later review, sharing, or progress tracking. JumpSense AI allows users to fine-tune the counter to match their personal style through in-app settings. These settings include minimum jump amplitude, signal smoothing, idle sensitivity, double-under interval thresholds, and extrema detection window. To accommodate different devices and heat management needs, the app supports multiple frame rates. Performance mode runs at 60 FPS for accurate double-under detection at high speeds. Balanced and Power Saver modes run at 30 or 15 FPS for older devices to reduce heat.

Use Cases

JumpSense AI is suitable for jump rope fitness and HIIT workouts. It is also great for double-under practice for CrossFit and fitness training. The app works well for boxing and MMA jump rope warm-ups. It is useful for school PE classes and jump rope teams. Home workouts without the need for smart ropes or fitness trackers are another common use case. To use the app, users open it and grant camera permission. They position the iPhone so the camera captures their full body. They start jumping and the count updates live. Users can tap Reset between sets or switch display modes as needed. Tapping Record saves a video with the overlay baked in.

Pricing

JumpSense AI is a one-time purchase that costs $9.99. This single payment unlocks all features. There are no recurring fees, in-app purchases, or hidden costs.

Vibes

JumpSense AI is an independent fitness utility that is not affiliated with Apple Inc. It uses Apple's publicly available Vision framework. The app is developed by Hu Yue and was released in 2025. The developer has made a clear privacy promise that nothing leaves the user's iPhone. There is no account creation, login, analytics, or data collection. Recorded videos remain in the user's Photos library. The app is compatible with iOS 17.0 or later, Macs with macOS 14.0 and an Apple M1 chip or later, and Apple Vision devices with visionOS 1.0 or later. The app is available in English and Simplified Chinese. The file size is 1.7 MB and the age rating is 4+.

Additional Information

Users should consult a physician before starting any exercise program. The app relies on the system framework VNDetectHumanBodyPoseRequest and uses a COCO-17 skeleton. The algorithm uses ankle-Y peak/valley detection with EMA smoothing and hip-baseline EMA compensation.