Introduction
VideoDraft integrates the most advanced AI models from leading providers to deliver exceptional video creation capabilities. Each model is carefully selected for specific strengths, ensuring you have the right tool for every creative need.Model availability varies by plan. Pro plans and above unlock access to premium models for superior quality.
Model Categories
Image Generation
Create stunning visuals with models from Flux, OpenAI, and Google
Video Generation
Generate dynamic videos with Luma, Kling, and Google Veo models
Speech Synthesis
Natural voiceovers with Google’s advanced text-to-speech
Quick Model Comparison
Image Models
| Model | Provider | Credits | Speed | Best For |
|---|---|---|---|---|
| Flux Schnell | Black Forest Labs | 5 | ⚡ Fast | Quick drafts |
| Flux Pro | Black Forest Labs | 5 | 🔄 Medium | Balanced quality |
| Flux Kontext Pro | Black Forest Labs | 5 | 🔄 Medium | Style matching |
| GPT-Image (DALL-E 3) | OpenAI | 10 | 🔄 Medium | Creative concepts |
| Imagen-4 | 10 | 🐌 Slower | Photorealism |
Video Models
| Model | Provider | Credits | Duration | Key Feature |
|---|---|---|---|---|
| Kling v2.1 | Kuaishou | 25 | 5 sec | Best value |
| Luma Ray Flash | Luma AI | 30 | 5 sec | Fast generation |
| Luma Ray 2 | Luma AI | 90 | 5 sec | High quality |
| Google Veo2 | 300-400 | 5-8 sec | Premium quality | |
| Google Veo3 | 600 | Variable | Integrated sound |
Speech Models
| Model | Provider | Credits | Languages | Features |
|---|---|---|---|---|
| Google TTS Wavenet | 1/100 chars | 50+ | Natural voices | |
| Google TTS Neural2 | 1/100 chars | 50+ | Enhanced quality |
Model Selection Guide
By Use Case
- Marketing Videos
- Educational Content
- Creative Projects
- High-Volume Production
Recommended Stack:
- Images: GPT-Image or Imagen-4
- Videos: Kling v2.1 or Veo2
- Speech: Google TTS Neural2
By Quality Requirements
Draft Quality- Use Flux Schnell for images
- Kling v2.1 for videos
- Fast, affordable iterations
- Flux Pro or GPT-Image
- Luma Ray 2 or Veo2
- Professional results
- Imagen-4 for photorealism
- Veo3 for cinematic videos
- No compromises
Model Capabilities
Image Generation Strengths
Flux Models
Flux Models
- Schnell: Lightning fast, good for iterations
- Pro: Excellent detail and composition
- Kontext Pro: Style consistency with references
- Best for general purpose imagery
GPT-Image (DALL-E 3)
GPT-Image (DALL-E 3)
- Creative interpretations
- Excellent with text in images
- Artistic and conceptual work
- Understands complex prompts
Google Imagen-4
Google Imagen-4
- Photorealistic results
- Accurate human representations
- Natural lighting and textures
- Professional photography quality
Video Generation Capabilities
Kling v2.1
Kling v2.1
- Reliable 5-second clips
- Good motion quality
- Affordable option
- Wide style range
Luma Ray Series
Luma Ray Series
- Ray Flash: Quick previews
- Ray 2: Superior quality
- Smooth motion
- Cinematic feel
Google Veo Series
Google Veo Series
- Veo2: Premium 4K quality
- Veo3: Integrated sound design
- Longest clips available
- Film-grade output
Credit Optimization Strategies
Smart Model Selection
-
Iterative Approach
- Start with low-cost models
- Refine prompts cheaply
- Switch to premium for finals
-
Model Matching
- Flux for illustrations
- Imagen for photos
- GPT-Image for concepts
-
Batch Processing
- Group similar requests
- Use consistent models
- Leverage caching
Credit-Saving Tips
For Images
- Test with Flux Schnell
- Use specific prompts
- Generate fewer variations
- Reuse successful prompts
For Videos
- Start with Kling v2.1
- Keep clips short
- Plan sequences carefully
- Use Veo3 sparingly
Technical Specifications
Resolution and Quality
Image Models:- Flux: Up to 1024x1024
- GPT-Image: Up to 1792x1024
- Imagen-4: Up to 2048x2048
- Kling: 720p
- Luma: 720p-1080p
- Veo2: Up to 4K
- Veo3: Variable HD/4K
Processing Times
Average generation times:- Images: 10-45 seconds
- Videos: 2-10 minutes
- Speech: Near instant
Processing times vary based on server load and model complexity. Premium models typically take longer but deliver superior results.
Model Updates and Improvements
Recent Enhancements
- Veo3 Launch: Integrated sound generation
- Flux Kontext Pro: Style reference support
- Imagen-4: Improved photorealism
- Kling v2.1: Better motion coherence
Coming Soon
- Enhanced character consistency
- Longer video generation
- Real-time previews
- Custom model training
Best Practices
Prompt Engineering
- Be Specific: Include details about style, lighting, composition
- Use References: Mention artistic styles or examples
- Technical Terms: Leverage photography/film terminology
- Iterate: Refine based on results
Model-Specific Tips
For Flux:- Keep prompts concise
- Focus on main elements
- Use style keywords
- Be creative and descriptive
- Include emotional context
- Experiment with concepts
- Include sound descriptions
- Use cinematic language
- Plan complete scenes
Ethical Considerations
Content Guidelines
All models follow strict ethical guidelines:- No harmful content
- Respect for individuals
- Copyright compliance
- Age-appropriate content
Usage Rights
- Commercial use allowed
- Attribution not required
- Model-specific restrictions apply
- Review terms for each model
Getting Support
Model Issues
If you experience problems:- Check credit balance
- Verify prompt compliance
- Try alternative models
- Contact support
Feature Requests
We continuously evaluate new models. Submit requests for:- Specific capabilities
- New providers
- Quality improvements
- Workflow integrations