- Allow multiple concurrent image generation requests
- Track pending batches with loading placeholders in gallery
- Show progress for each batch independently
- Warn before leaving page with pending generations
- Add clipboard paste support for reference images
- Paste images directly with Ctrl+V/Cmd+V
- Works from screenshots, copied images, etc.
- Add download button on image card hover
- Quick download without opening modal
- Timestamped filenames for organization
- Replace all emoji icons with clean SVG icons
- Professional Lucide/Feather-style icons
- Logo, buttons, loading states, empty states
- Consistent visual language throughout
- Persist image count setting across refreshes
- Save to localStorage on change
- Properly initialize input value on load
- Remove blocking loading indicator
- Generate button stays enabled during generation
- Each request shows inline loading placeholders
A client-side AI image generation tool using OpenRouter API.
Features:
- Multi-model support (Gemini, GPT-5, Flux 2, Seedream, Riverflow)
- Unlimited reference image uploads with drag & drop
- IndexedDB storage for persistent image gallery
- Adjustable resolution (1K/2K/4K) and aspect ratios
- Batch generation (up to 8 images)
- Custom styled dropdown for model selection
- Gallery with delete functionality on hover
- Full image modal with metadata and recreate option
- XSS protection with input sanitization
Tech Stack:
- Pure HTML/CSS/JavaScript (no dependencies)
- IndexedDB for image persistence
- OpenRouter API integration
License: GPL-3.0
Yusuf İpek