Embaixadinha
Football juggling practice, measured with on-device computer vision.
Project Overview
Embaixadinha is a native iOS app for tracking football juggling practice. Players record a session or import a video, then review their touches, streaks, and personal records. Video analysis runs on the device, and practice history is stored locally without requiring an account.
From Video to Practice Feedback
Core ML ball detection and Apple's Vision framework for body-pose tracking feed a pure-Swift analysis pipeline that identifies juggling touches. The app draws the results over the replay so players can inspect their session.
The same overlay renderer is shared by replay and export. An AVFoundation share editor supports trimming, pinch-and-pan framing, and exports in 9:16, 4:5, and 16:9 formats.
Key Features
- On-device analysis: ball detection, pose tracking, and touch counting without uploading practice videos.
- Replay and sharing: inspect touch overlays and export a framed, trimmed MP4.
- Practice history: sessions, streaks, and personal records persisted with SwiftData.
- Native product experience: SwiftUI screens, localized content, and RevenueCat subscriptions for Pro tools.
Screenshots
Tech Stack
- SwiftUI
- Swift 6
- Core ML
- Vision
- AVFoundation
- SwiftData
- RevenueCat
Have a useful problem to solve?
Tell me what you are making, where it is getting stuck, and what a good outcome looks like.