AI image generation platform
U-Plast Studio — An AI-powered image generation platform
Project overview
U-Plast Studio is a proprietary web application for generating images using AI, developed as an internal tool for the rapid creation of visual content.
The product allows users to transform a text-based idea and reference images into a finished visual in just a few steps — from concept description to downloading the result.
The main focus of the project is on a simple user workflow and a clean interface, without being overloaded with features.
Task:
Create a user-friendly and technologically advanced image generation tool that:
- simplifies the use of AI for end users
- delivers quick results without complex settings
- allows you to control access and manage users
- scales as a fully-fledged product
The problem:
Most AI tools are overloaded with settings, require training and are unsuitable for quick tasks:
- complex interfaces
- lack of a clear UX workflow
- no user control in a corporate environment
- fragmented solutions without a unified system
Solution:
UX and user flow
The product is based on a linear and intuitive flow:
- Authorisation
- Selection of generation parameters
- Entering a text description
- Adding references
- Image generation
- Viewing and downloading the result
Each action logically follows on from the previous one, without any unnecessary steps.
Interface and UI
The interface is designed in a minimalist AI style:
- a light, tech-inspired theme
- glassmorphism and blur effects
- a card-based layout
- smooth animations and micro-interactions
- full-screen results view
Particular attention has been paid to:
- the speed at which the interface is perceived
- visual clarity
- the absence of visual clutter
Features
For users:
- Generate images from a text prompt
- Support for resolutions (1K / 2K / 4K)
- Upload reference images
- View results in full screen mode
- Download the final image
For the administrator:
- User management
- Creating and deleting accounts
- Access control
- Viewing generation statistics
Technical implementation
Frontend:
- React + TypeScript
- Vite
- Tailwind CSS
- custom UI components
- React Hooks и Portals
Backend:
- Go (Chi router)
- JWT-login
- PostgreSQL
- integration with Gemini API
Infrastructure:
- Docker Compose
- Nginx (production)
- dev proxy через Vite
Architecture
The project is structured as follows:
- SPA frontend (interface and UX)
- API backend (logic, authorisation, generation)
The frontend communicates with the backend via a REST API, sending generation data and receiving the result in image format.
Result:
The result was the creation of a fully-fledged AI product MVP, which:
- is ready for use within the company
- demonstrates the full product lifecycle (UI → API → AI → result)
- can be scaled up to a production service
- can be expanded into a SaaS solution
Business value
- Faster creation of visual content
- Reduced reliance on designers during the concept phase
- Centralised access management
- A foundation for a commercial AI product
Development potential
- History of generations and gallery
- Pricing plans and limits
- Cloud image storage
- Advanced AI settings
- Teamwork and sharing
