Skip to content

Project CIG

Contorr Image Gen — AI human models and virtual try-on

Category: Product & client work

  • Built with:
  • Next.js 16 (App Router)
  • React 19
  • TypeScript
  • Tailwind CSS v4
  • Prisma
  • PostgreSQL
  • Redis
  • BullMQ
  • MinIO/S3
  • Fal.AI
  • ComfyUI
  • Express
  • Zod
  • argon2
  • Docker
  • Railway
  • pnpm
  • Vitest
The Human Models gallery — a grid of AI-generated model images, each stored with the parameters that produced it.

What it is

A full-stack app that generates hyper-realistic human models with AI, then dresses them in real apparel photos via virtual try-on. A guided prompt builder and per-model parameter forms drive every render, and each output is stored with its model and parameters.

Value

Replaces studio photo shoots for apparel catalogs: on-model product imagery across body types, skin tones and settings in minutes instead of days, at a fraction of the cost. Stored parameters make winning looks reproducible across a whole product line.

Technology highlights

  • Async AI pipeline — a BullMQ/Redis worker submits jobs to Fal.AI (or a self-hosted ComfyUI proxy), polls to completion and streams results into S3/MinIO, keeping the web tier responsive.
  • Type-safe pluggable model registry — declarative model specs render the parameter forms dynamically and are Zod-validated end to end, so new AI models need no UI changes.

More projects