Product Pics is an AI-powered product photo generator for apparel catalogs. It turns flat product shots into ecommerce-ready lifestyle images by placing age-appropriate models into the scene with varied natural backgrounds.
The app is built for batch catalog work, not one-off image editing. Upload up to 100 product images, choose a demographic and age range, run generation asynchronously, and download the full results as a zip archive.
The app compares the original source image against the generated lifestyle result inside each batch:
- Generates product photos for apparel catalogs using AI.
- Supports batch uploads of up to 100 source images.
- Lets users choose
baby,boy,girl,man, orwomandemographics plus a single age or age range. - Assigns diverse natural backgrounds across a batch for visual variety.
- Tracks progress while generation jobs run asynchronously.
- Packages completed outputs into a zip file with a summary report.
- Stores uploads and generated images in Cloudflare R2.
- Keeps per-user batch history behind authentication.
The higher-level product walkthrough lives in docs/feature-overview.md.
- Framework: Next.js 16 + React 19
- Authentication: better-auth
- Database: Postgres + Drizzle ORM
- Storage: Cloudflare R2
- Image generation: RunPod
nano-banana-edit - Image processing: Sharp
- Clone the repository:
git clone https://github.com/amxv/product-pics.git
cd product-pics- Install dependencies:
bun install- Create
.env.local:
BETTER_AUTH_SECRET=""
BETTER_AUTH_URL="http://localhost:3000"
NEXT_PUBLIC_BETTER_AUTH_URL="http://localhost:3000"
DATABASE_URL=""
R2_ACCOUNT_ID=""
R2_ACCESS_KEY_ID=""
R2_SECRET_ACCESS_KEY=""
R2_BUCKET_NAME=""
R2_PUBLIC_URL=""
RUNPOD_API_KEY=""- Start the app:
bun devOpen http://localhost:3000.
- User signup is disabled in the current product flow; accounts are expected to be provisioned separately.
tmp/is intentionally gitignored for local scratch assets and experiments.- The app currently focuses on generation, tracking, and download workflows rather than interactive retouching.
- The current UX is especially strong for apparel workflows where the source image needs to stay faithful to garment color, texture, and branding.
Apache License 2.0. See LICENSE.