Personalized AI Media App Development Swipe like Tinder

Project scope:
• Upload a few profile images.
• Discover AI-generated “personalized media” in a Tinder-style swipe interface.
• Save favorites
• Use a credit/paywall system to control generation volume. (Stripe)

A Node.js backend will orchestrate:
• Auth, credits, generation queueing with a third-party API for Image generations (HTTP API).
• Storage of user uploads and generated assets.
• An Admin Panel to manage item catalog, links, pricing packs, and monitor jobs.

2) Core User Flows
• Onboarding & Auth: Email/Google sign-in.
• Profile Upload: User uploads 2–4 images.
• Discovery Feed (Swipe UI): Card stack with infinite scroll; each card shows a generated asset and has actions: Save / Dismiss / Open Link.
• Favorites: Saved items (when swiped right)
• Credits & Paywall: Each generation consumes 1 credit. Users can purchase credit packs or subscribe.

3) Admin Panel (Required)

Role: admin
• Catalog Management
• CRUD items (title, brand/source, preview image URL, destination URL, tags, active flag).
• Bulk import (CSV).
• Link Management
• Configure outbound links per item (track clicks).
• Pricing
• Manage credit packs (e.g., 20/100/300), subscriptions (basic/premium), prices, and visibility.
• Users & Credits
• View users, credit balances, last activity.
• Manual credit adjustments; ban/unban.
• Jobs Monitor
• View generation jobs (queued/running/done/failed), latency, error logs.
• Reports
• Daily generation cost, credits spent, top items by saves/clicks.

(Admin can be a simple Next.js/React SPA behind auth or an integrated dashboard.)