Primeshot | AI Photo Studio

  • UX Design
  • UI Design
  • Art Direction
  • Figma
  • React

Role: Co-founder → primeshot.ai | 2025

Responsibilities:

As co-founder, I led the creation of the brand identity, UI design for both the app and the website, and the entire front-end implementation and product architecture.

Context:

Primeshot is an AI-powered photo studio that generates professional, creative, or editorial-style portraits in minutes from a handful of selfies. The mission: make high-quality photography accessible, instant, and fully customizable through a monthly subscription and a library of new styles released every week.

Branding

Visual Identity

Primeshot’s identity was crafted to feel memorable, energetic, and distinctly modern, reflecting a technology that is precise, advanced, and highly innovative. The system relies on a fresh palette (teal, turquoise, deep black, white) accented with orange, paired with a geometric symbol and a clean wordmark. Together, they communicate innovation, simplicity, and dynamism.

Art Direction

The visual identity borrows from photo-studio aesthetics and creative tooling: matte surfaces, controlled lighting, color blocks, and modular layouts. The direction remains intentionally understated to let generated portraits take center stage, while still giving the brand a strong, instantly recognizable presence. The goal: a professional, contemporary, and intuitive identity.

App Design

Process

Due to time constraints, we opted to skip a lengthy user-research phase.
Instead, we focused on:

  • A streamlined flow that moves quickly from onboarding to photo generation

  • High-fidelity screens from day one to speed up decision-making

  • An interface optimized for fast interaction: upload → choose style → generate

Design Principles

  • Ultra-clear UI with no unnecessary elements

  • Strong emphasis on imagery and styles

  • Consistent components powered by a shadcn/UI-based design system

  • Real-time feedback (training progress, generation status, queue position)

  • Accessibility baked into core components

Outcome

Users immediately understand:

  1. How to create their digital double

  2. How to browse and select photo styles

  3. How to launch a generation session

Development

Tech Stack

The product is built on a modern, high-performance, scalable architecture:

  • Front-end : Next.js 15, React 19, TypeScript, shadcn/UI
  • Back-end : Supabase (Auth, PostgreSQL, Edge Functions)
  • Stockage & CDN : AWS S3 + CloudFront
  • IA / GPU : Modal for LoRA training and image generation
  • Workflow IA : Optimized ComfyUI pipeline, 4K output, user-specific LoRA models
  • Paiements : Stripe (subscriptions + credit system)
  • Monitoring : Vercel Analytics, Sentry
Key Features Developed

  • Personal model creation and training in ~8 minutes

  • Ultra-high-resolution image generation in minutes

  • Subscription and credit-based billing

  • Secure upload pipeline with automated quality validation

  • Full dashboard for characters, styles, and generation history

  • Admin interface for publishing new styles

  • Serverless scalability (GPU fallback, queues, edge functions)

Conclusion

Although still early in its trajectory, we built:

  • A complete, scalable product

  • A strong brand in a crowded market

  • Proprietary technology with full control over the AI pipeline

  • A solid foundation for continuously releasing new styles and improving retention