Using Trained Models
Create videos with your custom trained models
Streamlined workflow for consistent results
Skip the image generation step and create videos directly with your trained people, products, or styles.
See Trained Models in Action
Watch how trained models streamline your video creation workflow and ensure consistent results.
What Are Trained Models?
Trained models are custom AI models created from your photos. Instead of generating an image first and then animating it, you can use trained models directly in video generation for faster, more consistent results.
Why Use Trained Models in Video?
50% Faster Workflow
Skip image generation step entirely
Perfect Consistency
Same face/product/style every time
Brand Control
Use your exact products and people
Cost Effective
One prompt instead of two separate generations
Traditional vs. Trained Model Workflow
See the difference in time and steps when using trained models in video generation.
Without Trained Models
8 steps • 5-8 minutes
Go to Image Generator
Write prompt with your subject
Generate and select best image
Download the image
Go to Video Generator
Upload the image
Write motion prompt
Generate video
With Trained Models
4 steps • 2-3 minutes
Go to Video Generator
Select your trained model
Write scene + motion prompt
Generate video directly
Time saved: 50%+ faster with consistent results
Types of Trained Models for Video
Different trained models work for different video use cases. Here's how to use each type effectively.
Person Models
Your face or specific people
Key Benefits:
- Consistent face across videos
- Professional content creation
- Personal branding
Best for: Create branded content, spokesperson videos, or personal vlogs with consistency
Example Prompt:
"Sarah_headshots model walking through a modern office, professional lighting, confident stride"
Product Models
Your products or brand items
Key Benefits:
- Brand consistency
- Multiple video angles
- Professional showcases
Best for: Generate product demos and marketing videos with your exact products
Example Prompt:
"luxury_watch model rotating on marble surface, studio lighting catches metal details, premium commercial feel"
Style Models
Specific visual aesthetics
Key Benefits:
- Consistent visual brand
- Artistic control
- Unique style
Best for: Maintain visual brand consistency across all your video content
Example Prompt:
"vintage_film model showing bustling city street, warm nostalgic tones, classic cinematography"
How to Use Trained Models in Video Generation
Select Your Trained Model
Choose from your existing models in the video generator
- Browse your trained models library
- Preview model examples to verify quality
- Select the model that fits your video concept
Write Your Prompt
Combine your model name with scene and motion description
- Reference your model: [model_name] doing [action]
- Add setting: in [environment/location]
- Include camera and lighting directions
Configure & Generate
Set your video parameters and create the video
- Choose duration (5 or 10 seconds)
- Select quality level based on your needs
- Generate and review the result
Tips for Best Results
Best Practices
Use exact model name as it appears in your library
Describe environments and actions clearly
Keep one main model per video for consistency
Pro Tips
Test prompts with medium quality first to save credits
Use same prompting techniques as text-to-video
Preview model examples to match your video vision
Ready to Use Your Trained Models?
Now you understand how to streamline your video workflow with trained models. Select your best model and start creating consistent video content.