Contact person: Yoga Learning App with AI Scoring -- 2

Phone:Show

Email:Show

Location: Bengaluru, India

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"I need a straightforward cross-platform mobile app—one build that runs smoothly on both iOS and Android—that mirrors the essentials of WHO’s mYoga while adding an AI-driven posture score.

Core flow
• A new user lands on a minimal sign-up screen and can register with either a mobile number (OTP) or an email address (verification link). During onboarding I must capture name, age, and whether the user is at a basic or advanced experience level.
• After sign-up the app surfaces ten pre-loaded yoga videos: five tagged “basic”, five tagged “advanced”. Only the set that matches the user’s selected level appears in the initial playlist, but the other category should remain accessible from a simple toggle.
• Users tap a video to watch and learn, then switch to “Practice” mode. Here they can snap or upload a photo of their pose and press an AI Score button. Your code should call an embedded or cloud model (TensorFlow Lite / MediaPipe Pose, or an equivalent public checkpoint) to compare the user’s posture with the ideal frame and return a similarity score on screen. No custom model training is required beyond integrating the existing open-source weights.

Deliverables
• React Native or Flutter codebase (open to either) compiled and tested for iOS & Android
• Simple, clean UI matching the flow above
• Lightweight backend or local storage to keep the ten videos and basic profile data (Firebase, Supabase, or SQLite—whichever is faster for you)
• AI posture scoring module wired into the Practice screen, returning a numeric score plus a brief text assessment
• Build & deployment instructions so I can generate new releases myself

Acceptance criteria
1. A fresh install lets me register through mobile or email, fills name/age/level once, and lands on the correct playlist.
2. All ten videos play without buffering in normal network conditions.
3. Tapping AI Score returns a similarity score in under five seconds on a midrange phone.
4. The same APK/IPA passes store review requirements.

If this scope is clear and you’ve shipped pose-estimation or fitness apps before, let’s talk timelines and milestones." (client-provided description)