What You’ll Learn
AI Visual Generation
Generate unique images with AI instead of stock visuals
Custom Backgrounds
Create visuals from text prompts automatically
Multiple AI Models
Choose from different AI models and media styles
Unique Content
Stand out with one-of-a-kind visuals
Before You Begin
Make sure you have:- A Pictory API key (get one here)
- Node.js or Python installed on your machine
- An idea of the visual style you want to create
- Basic understanding of AI image generation concepts
How AI Visual Generation Works
When you use AI-generated visuals in Pictory:- Prompt Processing - Your text prompt is analyzed to understand the desired image
- AI Generation - The selected AI model creates a unique image based on your prompt
- Style Application - Media style preferences are applied to match your brand
- Video Integration - The generated image is seamlessly integrated as your scene background
- Automatic or Manual - Use custom prompts or let AI generate prompts from your story
AI visual generation takes longer than using stock visuals because each image is created specifically for your video. Plan for additional processing time when using this feature.
Complete Example
Understanding the Configuration
Background Object
| Parameter | Type | Required | Description |
|---|---|---|---|
type | string | Yes | Must be set to "image" for AI-generated visuals |
aiVisual | object | Yes | Configuration for AI visual generation |
AI Visual Parameters
| Parameter | Type | Required | Description |
|---|---|---|---|
prompt | string | No | Text description of the visual to generate (max 250 characters). If omitted, AI generates prompts from your story |
model | string | Yes | The AI model to use for image generation |
mediaStyle | string | No | The visual style to apply to the generated image |
Available AI Models
| Model | Speed | Quality | Best Used For |
|---|---|---|---|
flux-schnell | Fast | Good | Quick iterations, testing, drafts, time-sensitive content |
seedream3.0 | Medium | Balanced | General-purpose content, versatile applications |
nanobanana | Medium | Specialized | Specific artistic styles, branded content |
titan | Slower | Excellent | Final production, high-quality marketing, premium content |
Available Media Styles
| Style | Visual Characteristics | Best Used For |
|---|---|---|
photorealistic | Realistic photographs, natural lighting | Corporate videos, professional presentations, realistic scenarios |
artistic | Artistic renderings, painterly effects | Creative content, brand storytelling, abstract concepts |
cartoon | Cartoon-style images, bold colors | Children’s content, educational videos, fun marketing |
minimalist | Simple, clean designs, reduced details | Modern branding, tech content, professional minimalism |
vintage | Retro aesthetic, aged appearance | Nostalgia marketing, historical content, unique branding |
futuristic | Modern, sci-fi look, high-tech feel | Technology content, innovation topics, forward-thinking brands |
Auto-Generated Prompts
You can omit theprompt field to let the AI automatically generate prompts based on your scene’s story text:
When to Use Auto-Generated Prompts:
- Creating videos with multiple scenes (
createSceneOnEndOfSentence: true) - You want the AI to interpret your story text visually
- Quick content creation without manual prompt writing
- Testing different visual interpretations
Common Use Cases
Technology and Innovation Content
Educational and Learning Content
Marketing and Brand Content
Historical or Vintage Content
Best Practices
Write Effective Prompts
Write Effective Prompts
Craft prompts that generate the visuals you need:
- Be specific: Include details about composition, lighting, and mood
- Use descriptive language: “bright morning sunlight” vs “sunny”
- Mention key elements: “modern office with glass walls and city view”
- Keep under 250 characters: Concise prompts work better
- Avoid negatives: Say what you want, not what you don’t want
Choose the Right Model
Choose the Right Model
Match your AI model to your use case:
- Testing phase: Use
flux-schnellfor quick iterations - Production: Switch to
titanfor final, high-quality output - General content:
seedream3.0offers good balance - Budget conscious:
flux-schnellprovides faster, more economical results
Select Appropriate Media Styles
Select Appropriate Media Styles
Match style to your brand and content:
- Corporate/Professional: Use
photorealisticorminimalist - Creative/Artistic: Try
artisticorvintage - Tech/Innovation: Go with
futuristic - Educational/Fun: Consider
cartoon - Consistent branding: Stick to one style across your videos
Plan for Processing Time
Plan for Processing Time
AI visual generation takes longer than stock visuals:
- Expect 5-15 minutes for generation depending on model
flux-schnellis fastest (2-5 minutes per image)titantakes longer but produces premium quality (10-15 minutes)- Test with faster models before final production
- Don’t use AI generation for urgent, time-critical videos
Iterate and Refine
Iterate and Refine
Perfect your visuals through iteration:
- Start with auto-generated prompts to see AI interpretation
- Refine prompts based on initial results
- Test different models and styles
- Save successful prompts for reuse
- Document what works for your brand
Troubleshooting
Generated image doesn't match prompt
Generated image doesn't match prompt
Problem: The AI created an image that doesn’t reflect your prompt description.Solution:
- Make your prompt more specific and detailed
- Add descriptive adjectives (e.g., “bright”, “modern”, “spacious”)
- Specify composition: “aerial view”, “close-up”, “wide angle”
- Include lighting details: “sunset lighting”, “studio lighting”
- Try a different AI model - some interpret prompts differently
- Example revision:
- Before: “office”
- After: “modern open-plan office with glass walls, natural daylight, and minimalist furniture”
Image quality is poor
Image quality is poor
Problem: Generated images look blurry, pixelated, or low quality.Solution:
- Switch to a higher quality model:
- Replace
flux-schnellwithtitan - Use
seedream3.0for balanced quality
- Replace
- Avoid very complex prompts (keep under 250 characters)
- Try a different media style that suits the content better
- Ensure your prompt describes a clear, achievable visual
Video processing takes too long
Video processing takes too long
Problem: Job status shows “in-progress” for extended periods.Solution:
- AI visual generation is slower than stock visuals - this is normal
- Expected times:
flux-schnell: 5-8 minutesseedream3.0: 8-12 minutestitan: 12-20 minutes
- Multiple scenes with AI visuals will multiply processing time
- Consider using AI visuals only for key scenes
- Use stock visuals for time-sensitive projects
Error: Invalid background configuration
Error: Invalid background configuration
Problem: API returns an error about background settings.Solution:
Check your background configuration:
AI generated unexpected content
AI generated unexpected content
Problem: The image contains elements you didn’t want.Solution:
- Be more specific about what you want to see
- Add constraints to your prompt: “simple”, “minimal”, “clean”
- Avoid ambiguous language that AI might misinterpret
- Test prompts iteratively to refine results
- Focus prompts on primary subject matter only
Next Steps
Enhance your AI-generated visual videos with these features:AI Voice-Over
Add professional narration to your videos
Brand Settings
Apply consistent branding automatically
Custom Captions
Add translated or custom subtitles
Background Music
Add music to complement your visuals
