Overview
AI Studio is VideoDraft’s creative playground where you can generate individual assets using cutting-edge AI models. Whether you need a specific image, video clip, or sound effect, AI Studio provides direct access to the best AI generation tools.AI Studio is perfect for creating custom assets, experimenting with different models, or generating content outside of the workflow system.
Available Generators
- Image Generation
- Video Generation
- Sound Generation
Create stunning images with multiple AI models:
- Flux Models (Schnell, Pro, Kontext)
- GPT-Image (DALL-E 3)
- Google Imagen-4
- Various styles and resolutions
Image Generation
Available Models
Flux Schnell
5 credits per image
- Fastest generation
- Good quality
- Best for drafts
Flux Pro
5 credits per image
- High quality
- Better details
- Professional results
Flux Kontext Pro
5 credits per image
- Context-aware
- Reference image support
- Style matching
GPT-Image (DALL-E)
10 credits per image
- Creative interpretations
- Artistic styles
- Text rendering
Google Imagen-4
10 credits per image
- Photorealistic
- Accurate details
- Natural lighting
Image Generation Tips
-
Prompt Structure
-
Example Prompts
- “Professional headshot of a confident businesswoman, studio lighting, shallow depth of field, corporate style”
- “Majestic mountain landscape at golden hour, dramatic clouds, cinematic composition, 4K quality”
- “Futuristic city skyline at night, neon lights, cyberpunk aesthetic, aerial view, ultra detailed”
-
Style Keywords
- Photographic: “DSLR, 85mm lens, bokeh, RAW”
- Artistic: “oil painting, watercolor, digital art”
- Technical: “4K, ultra HD, highly detailed, sharp”
Advanced Image Features
Aspect Ratios
Aspect Ratios
- 1:1 - Square (Instagram posts)
- 16:9 - Landscape (YouTube, presentations)
- 9:16 - Portrait (Stories, Reels)
- 4:3 - Classic (Traditional photos)
- 3:2 - Photography standard
Style Presets
Style Presets
- Photorealistic
- Artistic/Painterly
- Anime/Manga
- 3D Render
- Minimalist
- Abstract
Reference Images
Reference Images
With Flux Kontext Pro:
- Upload a reference image
- AI matches style and composition
- Maintain brand consistency
- Create variations
Batch Generation
Batch Generation
- Generate multiple variations
- Test different prompts
- Create image sets
- A/B testing
Video Generation
Video Models Comparison
| Model | Credits | Duration | Features | Best For |
|---|---|---|---|---|
| Luma Ray Flash | 30 | 5 sec | Fast, smooth | Quick previews |
| Luma Ray 2 | 90 | 5 sec | High quality | Final videos |
| Kling v2.1 | 25 | 5 sec | Good balance | Most projects |
| Google Veo2 | 300-400 | 5-8 sec | Premium quality | High-end content |
| Google Veo3 | 600 | Variable | Integrated sound | Complete videos |
Video Prompt Strategies
-
Motion Description
- Specify camera movements
- Describe action clearly
- Include speed/pacing
-
Cinematic Prompts
-
Technical Terms
- Pan, tilt, zoom, track
- Shallow/deep focus
- Time-lapse, slow-motion
- Aerial, low angle, bird’s eye
Video Generation Tips
- For Beginners
- For Quality
- For Efficiency
- Start with Kling v2.1
- Keep prompts simple
- Focus on one action
- Test short clips first
Sound Generation
Available Options
- Sound Effects: Specific sounds (footsteps, doors, nature)
- Ambient Audio: Background atmospheres
- Musical Elements: Short musical phrases
- Foley: Everyday sound effects
Sound Prompt Examples
Using AI Studio Effectively
Workflow Integration
-
Asset Creation
- Generate custom assets for projects
- Build libraries of reusable content
- Create brand-specific visuals
-
Experimentation
- Test different models
- Compare outputs
- Find your preferred style
-
Quality Control
- Preview before using in projects
- Generate alternatives
- Select best results
Organization Tips
Naming Convention
Use descriptive names:
- project_scene_version
- style_subject_date
- client_asset_type
Tagging System
Tag your generations:
- By project
- By style
- By model used
- By quality rating
History & Management
Generation History
- All generations are saved
- Searchable by prompt
- Filterable by type
- Downloadable anytime
Storage & Organization
- Unlimited storage for generated content
- Organize by projects
- Create collections
- Share with team (coming soon)
Best Practices
-
Start with Lower-Cost Models
- Test concepts with Flux Schnell
- Refine prompts cheaply
- Use premium models for finals
-
Iterate Intelligently
- Make small prompt adjustments
- Learn what each model excels at
- Build a prompt library
-
Credit Management
- Monitor usage in real-time
- Plan generations ahead
- Batch similar requests
-
Quality Over Quantity
- One great image > ten mediocre ones
- Spend time on prompt crafting
- Use reference images when possible
Model Selection Guide
Choose Flux Models When:
- You need fast results
- Working on drafts
- Creating multiple variations
- Budget is a concern
Choose GPT-Image When:
- You need creative interpretations
- Working with text in images
- Want artistic styles
- Creating unique concepts
Choose Imagen-4 When:
- Photorealism is critical
- Need accurate human faces
- Creating professional photos
- Quality is paramount
Choose Video Models Based On:
- Kling: Best value for most uses
- Luma Ray: When quality matters more
- Veo2: For premium productions
- Veo3: When you need integrated sound
Troubleshooting
-
Generation Failed
- Simplify prompt
- Remove prohibited content
- Check credit balance
- Try different model
-
Poor Quality Results
- Add more details
- Use quality keywords
- Try different model
- Adjust parameters
-
Not What Expected
- Review prompt clarity
- Add negative prompts
- Use reference images
- Break complex ideas down
Advanced Techniques
Prompt Engineering
- Use weight modifiers: (important element:1.2)
- Negative prompts: “no text, no watermarks”
- Style mixing: “in the style of [artist/movement]”
- Technical specs: “shot on Canon 5D, 85mm lens”
Consistency Tricks
- Save successful prompts
- Use seed values for variations
- Maintain style descriptors
- Create prompt templates
Next Steps
- Open AI Studio - Start creating
- Prompt Library - Example prompts
- Model Gallery - See what’s possible