Overview
The Media Generator allows you to create AI-generated images and videos using various models and settings. The interface adapts based on whether you're creating photos or videos, providing relevant options for each mode.
Accessing the Generator
1. Go to Tools → Media Generator
2. Click "Create Job* from the dashboard
3. The generator page opens with configuration sideba
Mode Selection: Photo vs Video
Toggle Tabs
At the top of the sidebar, you'll see two tabs:
Photo Tab (Camera icon)
- Generate static images
- Edit existing images
- Upscale images
- Remove backgrounds
Video Tab (Video camera icon)
- Generate videos from images
- Create video sequences
- Stitch multiple videos together
Switching Modes
- Click the tab you want to use
- Sidebar options update automatically
- Settings are preserved when switching
Photo Mode Options
Reference Images
Purpose: Images used as reference for generation
Adding References
- Upload: Click "Upload" to select images from your computer
- Your Products: Click "Your Products" to browse your product library
- Supports multiple images (for some models)
- Single image required for editing/upscaling models
Display
- Selected images shown in grid
- Click X to remove
- Required for certain models (indicated by warning)
Model Selection
Available Models
SORA
- Artistic results
- Lower consistency
- 0.8 credits per image
GPT Image 1.5
- OpenAI's latest model
- Improved efficiency
- 0.16 credits per image
Nano Banana PRO
- Text-friendly
- Strong consistency
- Gemini 3 Pro powered
- 0.22 credits per image
Qwen Image Edit
- Fast editing
- Consistent results
- Requires reference image
- 0.6 credits per image
Pruna Image Edit
- Fast, precise editing
- 0.052 credits per image
- Requires reference image
Seedream 4
- Great for product & lifestyle images
- 0.5 credits per image
Crystal Upscaler
- High-precision upscaling
- Requires reference image
- 0.6 credits per image
Remove Background
- Cuts subject from background
- Requires reference image
- 0.5 credits per image
Selecting a Model
- Click model dropdown
- View model descriptions and pricing
- Select from list
- Cost updates automatically
Aspect Ratio
Options
- Square (1:1): Perfect for social media
- Landscape (16:9): Wide format
- Portrait (9:16): Vertical format
- Standard (4:3): Traditional format
- Photo (3:4): Portrait photo format
Selection
- Click desired aspect ratio icon
- Selected ratio highlighted
- Not available for editing/upscaling models
Quantity
Slider Control
- Range: 1 to 12 images
- Drag slider or use input
- Cost multiplies by quantity
- Shows current count badge
Best Practices
- Start with 1-2 for testing
- Use more for variations
- Consider cost implications
Styles
Style Library
- Browse curated style presets
- Search by keywords
- Filter by categories
- Select up to 5 styles
Using Styles
- Click "Browse Library" button
- Search or filter styles
- Click style cards to select
- Selected styles shown as chips
- Remove by clicking X on chip
Style Categories
- Various categories available
- Filter sidebar shows counts
- Multiple categories can be selected
Prompt Canvas
Main Prompt
- Large text area for description
- Describe what you want to create
- Max 1800 characters
- Character counter shown on hover
Generate Unique Variations
- Checkbox option
- Creates unique prompts for each image
- Combines with styles automatically
- Useful for batch generation
Enhanced Prompts
- Auto-generated when using styles/variations
- Shows below main prompt
- Click to expand/collapse
- Remove individual prompts if needed
Special Model Options
Crystal Upscaler Settings
- Scale Factor: 1-10× upscale
- Creativity: 0-10 slider
- Output Format: PNG (recommended) or JPG
Remove Background Settings
- Preserve Partial Alpha: Keeps soft edges
- Better for hair, fur, transparent objects
Video Mode Options
Reference Frames
First Frame
- Starting image for video
- Required for all videos
- Upload or select from products
Last Frame (if supported)
- Ending image for video
- Only available for certain models
- Creates transition from first to last
- Optional but recommended
Adding Frames
- Upload: Upload new images
- Your Products: Select from library
- Dropdown menus for frame selection
- Clear buttons to remove frames
Video Model Selection
Available Models
MiniMax Hailuo v2.0
- 0.7 credits per video
- Supports first and last frame
Google Veo 3.1
- Dynamic pricing (0.75 × duration)
- Supports first and last frame
Seedance 1 Pro
- 1.75 credits per video
Luma Ray 2
- 1.9 credits per video
Kling v2.1
- 1.8 credits per video
- Supports first and last frame
Selecting Model
- Dropdown shows model logos
- View pricing information
- Select based on needs and budget
Video Settings
Duration
- Video length in seconds
- Typically 6 seconds default
- Affects cost for some models
Resolution
- Video quality setting
- Options like "768P"
- Higher resolution = better quality
Loop
- Checkbox option
- Makes video loop seamlessly
- Useful for product showcases
Video Count
- Number of videos to generate
- Typically 1-5 videos
- Cost multiplies by count
Video Sequences
Sequential Videos
- Create multiple connected videos
- Each sequence can have different settings
- Automatically links together
- Option to stitch into one final video
Sequence Configuration
- Add multiple sequences
- Each has first/last frame options
- Can use previous video's last frame
- Independent settings per sequence
Stitching
- Combines all sequences into one video
- Enabled via checkbox
- Creates seamless final output
Results Display
Job Results
Main Canvas Area
- Shows all generated jobs
- Scrollable list
- Newest jobs appear at top
Filtering Results
Type Filters
- All: Shows everything
- Images: Photo generation results
- Videos: Video generation results
- Edits: Image editing/upscaling results
Date Sorting
- Newest: Most recent first
- Oldest: Oldest first
Filter Behavior
- Click filter buttons
- Results update immediately
- Works with infinite scroll
Cost Estimation
Display
Estimated Cost
- Shown at bottom of sidebar
- Updates as you change settings
- Includes model cost × quantity
- For videos: includes duration pricing
Cost Factors
Photo Mode
- Model cost per image
- Quantity multiplier
- Example: 0.22 credits × 5 images = 1.1 credits
Video Mode
- Model base cost
- Duration multiplier (for some models)
- Sequence costs (if using)
- Example: 0.7 credits × 2 videos = 1.4 credits
Generation Process
Starting Generation
Generate Button
- Located at bottom of sidebar
- Shows estimated cost
- Disabled if:
- No job name entered
- Required reference missing
- Currently generating
Generation Steps
1. Prompt Generation (if enabled)
- Shows progress: "Generating prompt X/Y"
- Creates unique prompts with styles
- Only for image mode
2. Image/Video Generation
- Shows "Generating..." status
- Progress updates in real-time
- Results appear as they complete
During Generation
Status Indicators
- Loading spinner on button
- Progress updates
- Can continue working
- Results appear automatically
Job Creation
- Job appears in results area
- Scrolls to new job automatically
- Shows preview when available
- Real-time progress updates
Tips and Best Practices
Photo Generation
Reference Images
- Use high-quality references
- Multiple images improve results
- Ensure images match your vision
Prompts
- Be specific and descriptive
- Include style keywords
- Use "Generate unique variations" for diversity
Models
- SORA: Artistic, creative work
- Nano Banana PRO: Text-heavy content
- Seedream 4: Product photography
- Qwen/Pruna: Quick edits
Styles
- Combine 2-3 styles for best results
- Experiment with different combinations
- Use style library search
Video Generation
Reference Frames
- Use clear, high-quality images
- First frame sets the scene
- Last frame creates smooth transition
- Both frames improve results
Sequences
- Plan your sequence flow
- Use "previous" option for continuity
- Enable stitching for final video
Settings
- Longer duration = more cost (some models)
- Higher resolution = better quality
- Loop option for product showcases
Cost Management
- Check cost estimate before generating
- Start with smaller quantities
- Test with single images/videos first
- Use lower-cost models for experiments
Common Workflows
Creating Product Images
1. Switch to Photo mode
2. Upload product reference images
3. Select Seedream 4 or Nano Banana PRO
4. Choose aspect ratio (1:1 for social media)
5. Select relevant styles
6. Write descriptive prompt
7. Set quantity (1-3 for testing)
8. Click Generate
Editing Existing Images
1. Switch to Photo mode
2. Select reference image (required)
3. Choose*Qwen Image Edit or Pruna Image Edit
4. Write edit instructions in prompt
5. Click Edit IMAGE
Upscaling Images
1. Switch to Photo mode
2. Select image to upscale (required)
3. Choose Crystal Upscaler
4. Set scale factor (2-10×)
5. Adjust creativity slider
6. Choose output format
7. Click Upscale IMAGE
Creating Videos
1. Switch to Video mode
2. Select first frame image
3. Optionally select last frame
4. Choose video model
5. Set duration and resolution
6. Write motion description in prompt
7. Enable loop if needed
8. Click*Generate Video
Creating Video Sequences
1. Switch to Video mode
2. Set up main video (first sequence)
3. Add additional sequences
4. Configure each sequence's frames
5. Enable "Stitch sequences" if desired
6. Click Generate Video
Troubleshooting
Generation Fails
- Check reference images are selected (if required)
- Verify job name is entered
- Ensure sufficient credits
- Check model requirements
Results Not Appearing
- Wait for generation to complete
- Check job status in results
- Refresh page if needed
- Verify generation started successfully
Cost Higher Than Expected
- Check quantity settings
- Verify model pricing
- Review video duration (affects some models)
- Check sequence count for videos
Summary
The Media Generator provides powerful tools for creating AI-generated content:
Photo Mode: Generate, edit, upscale, and enhance images with various models, styles, and settings.
Video Mode: Create videos from images with control over frames, sequences, and output quality.
Key Features:
- Mode toggle (Photo/Video)
- Reference image management
- Model selection with pricing
- Style library integration
- Prompt canvas for descriptions
- Cost estimation
- Real-time results display
Use the mode tabs to switch between photo and video generation, configure your settings, and generate high-quality AI content tailored to your needs.