Project Overview
Objective
A full-stack AI image generation platform supporting multiple SDXL-based diffusion models on RunPod serverless GPU. Users can write prompts in any language — the system auto-translates to English — then configure generation parameters and generate images stored on AWS S3. Includes image inpainting, public gallery, per-user history, and an admin panel for managing users and images.
Stack
Next.jsTypeScriptNestJSFastAPIPrismaSQLiteAWS S3RunPod Serverless GPUPyTorchDiffusersSDXL TurboJuggernaut XLRealVisXL V5CyberRealistic XLLibreTranslateDockerJenkinsGoogle Analytics 4Sentry
Delivery highlights
- Supports 4 high-quality diffusion models: SDXL Turbo, Juggernaut XL, RealVisXL V5, and CyberRealistic XL, allowing users to select the best model for their creative intent.
- Automatic prompt translation from any language to English using LibreTranslate to ensure consistent high-quality embeddings across languages.
- Configurable generation parameters including image dimensions, seed, guidance scale, and steps for fine-tuned control over output.
- Full-featured image inpainting capability for editing and regenerating specific regions of generated or uploaded images.
- Public gallery showcasing top user-generated content with filtering and pagination for community discovery.
- Per-user image history with filtering and deletion options, and admin panel for user and content management.
- AWS S3 storage for persistent image retention with public CDN URLs for fast delivery and sharing.
- Real-time processing status updates via WebSocket-style communication and comprehensive error logging with Sentry.