Android Photo Enhancement MVP

I’m building a lean Android proof-of-concept that upscales and enhances photos with AI, then lets the user keep or share the result.

Core user flow
1. First-time launch: the user enters a phone number, receives an SMS OTP, and confirms it.
2. After verification, the user chooses an image either from the device gallery or by importing from a cloud drive via the Storage Access Framework.
3. The photo is uploaded to an AI enhancement endpoint, processed, and the improved version is returned with progress feedback.
4. The user can immediately download the file to local storage or share it through the native Android share sheet.

Mandatory implementation details
• Kotlin (preferred) or Java using a clean, easily extensible MVP/MVVM structure.
• All sensitive actions (storage, network, phone state if needed for OTP) must rely on Android’s runtime permission prompts, requested only at the precise moment of use.
• SMS OTP is the sole verification method; the backend you build should handle token generation, expiry, resend limits, and basic user bookkeeping.
• Integrate a single AI-based enhancement technique (e.g., Real-ESRGAN, Remini API, or a custom model you host) – no other filters are required for this milestone.
• Include a lightweight admin panel (even a simple web page) that lists users, their processed image count, and allows manual OTP resend.

Deliverables
• Full Android Studio project, well-commented, ready to compile on API 26+.
• Backend code (Node, Python, or similar) with README covering deployment and API keys for SMS and AI services.
• Admin panel source.
• A short setup guide plus test credentials so I can walk through the flow end-to-end.

Success is a build I can sideload, verify by SMS, pick or import a photo, watch it get AI-enhanced, and save or share— all while every permission request is transparent and confined to its stated purpose.