The official Android mobile client for WaveSpeed AI, built with React + Capacitor. Shares ~70% of code with the desktop version while being deeply optimized for mobile experiences.
- Current Version: 0.8.1
- Package Name:
ai.wavespeed.mobile - Minimum Android: 5.0 (API 21)
- Browse all available WaveSpeed AI models
- Search and filter support
- Tap any model to use it in Playground
- Input/Output Views: Mobile-optimized tab switching design
- Dynamic Forms: Auto-generated parameter forms based on model schema
- Supported Input Types:
- Text input (with AI prompt optimization)
- Sliders
- Dropdown selects
- Toggle switches
- File uploads (image/video/audio)
- LoRA selector
- Size selector
- Mask editor
- Real-time Pricing: Shows estimated cost
- Auto-switch: Automatically switches to output view when generation completes
- Save frequently used parameter configurations
- Rename templates
- Export/import templates (JSON format)
- Browse by model groups
- One-tap to apply template to Playground
- Grid view of generation history
- Filter by status: All, Completed, Failed, Processing, Created, Archived
- Thumbnail previews (image/video/audio/JSON/text)
- View detailed information
- Save as template
- Local input parameter caching
- API key management
- Account balance inquiry
- Theme switching (Auto/Dark/Light)
- Language selection (18 languages supported)
- Auto-save settings
Local AI tools that work without an API key:
- Frame-by-frame AI super-resolution upscaling
- 2x-4x upscaling options
- Three quality options (Fast/Balanced/High Quality)
- Real-time progress and ETA
- WebM output (30 FPS)
- AI image super-resolution
- 2x-4x upscaling
- ESRGAN models (slim/medium/thick)
- Download as PNG/WebP
- AI-powered automatic background removal
- Outputs three results simultaneously:
- Foreground (transparent background)
- Background (subject removed)
- Mask (grayscale segmentation)
- Auto-detects GPU acceleration
- Paint to remove objects from images
- LaMa inpainting model
- Brush/Eraser/Fill tools
- Undo/Redo support
- Smart crop for large image optimization
- Tap to select objects for segmentation
- Long-press to mark exclusion areas
- Real-time segmentation preview
- Feathered edge processing
- Multiple download formats
- Video format conversion
- Supports WebM (VP8/VP9/AV1) and MP4 (H.264)
- Auto codec detection
- Progress display
- Download the APK file
- Open the file on your phone
- If prompted about "unknown sources", allow installation in settings
- Open the app after installation
-
Get an API Key
- Visit WaveSpeed AI to register
- Get your API key from the user dashboard
-
Login
- Open the app and enter your API key
- Tap "Validate" button
- You'll be redirected to the main screen upon success
-
Using Playground
- Select a model from the Models page
- Configure parameters in the Input view
- Tap "Run" to start generation
- Auto-switches to Output view when complete
-
Save a Template
- Configure parameters in Playground
- Tap "Save as Template"
- Enter a name and save
-
Apply a Template
- Go to Templates page
- Tap the template you want to use
- Automatically redirects to Playground with loaded parameters
-
Export/Import
- Tap the export button at the top
- Choose to export single or all templates
- Import with merge or replace options
- Frontend: React 18 + TypeScript
- Mobile Framework: Capacitor 6.2
- UI Components: shadcn/ui + Tailwind CSS
- State Management: Zustand
- Build Tool: Vite
- AI Inference: ONNX Runtime, TensorFlow.js
- Node.js 18+
- Android Studio (for Android builds)
- Java 17+
# Install dependencies
npm install
# Start development server
npm run dev
# Build for production
npm run build
# Sync to Android
npx cap sync android
# Open Android Studio
npx cap open android
# Build Debug APK
npm run android:build:debug
# Build Release APK
npm run android:build:release
# Build Release Bundle (AAB)
npm run android:build:bundlewavespeed-mobile/
├── src/
│ ├── components/ # Mobile-specific components
│ │ ├── layout/ # Navigation, header, layout
│ │ └── playground/ # Form components
│ ├── pages/ # Mobile pages
│ ├── stores/ # State management
│ ├── hooks/ # Custom React hooks
│ ├── workers/ # Web Workers for AI processing
│ ├── platform/ # Capacitor platform services
│ └── App.tsx # Route configuration
├── android/ # Android native project
├── capacitor.config.ts # Capacitor configuration
├── vite.config.ts # Vite configuration
└── package.json # Dependencies
-
First-time AI Tool Usage
- Free tools download models on first use
- Models are cached locally for faster subsequent use
- Recommended to use Wi-Fi for initial downloads
-
Memory Usage
- AI tools have high memory consumption during processing
- Close other apps for smoother performance
- Large files may require more processing time
-
GPU Acceleration
- Some tools support WebGPU acceleration
- Automatically falls back to CPU when unavailable
- CPU mode is slower but more compatible
-
Video Processing
- Video enhancement takes considerable time
- Recommended for short videos (< 30 seconds)
- Output fixed at 30 FPS
-
Image Size
- Some tools have image size limits
- Large images are automatically scaled
- Recommended resolution: 4K or below
| Feature | Desktop | Mobile |
|---|---|---|
| Framework | Electron | Capacitor (Android) |
| Navigation | Sidebar | Bottom Navigation |
| Playground | Multi-tab | Single page + Input/Output toggle |
| Free Tools | Route-based | Persistent rendering |
| File Storage | electron-store | Capacitor Preferences |
| Drag & Drop | Full support | File picker |
| Video Converter | No | Yes |
- Input/Output view switching for better screen utilization
- Long-press interaction for Segment Anything exclusion
- Local input parameter caching in history
- Video converter tool
- Persistent free tools state across navigation
- Website: wavespeed.ai
- Issues: GitHub Issues
- API Docs: wavespeed.ai/docs
MIT License - see LICENSE for details.
