Skip to main content

Stable Video Diffusion - Image to Video

Generate high-quality videos from static images using StabilityAI's Stable Video Diffusion model with precise control over generation parameters.

Model Information​

  • Model ID: stable-video-diffusion
  • Display Name: Stable Video Diffusion - StabilityAI
  • Provider: StabilityAI

API Usage​

API Playground

https://api.1min.ai/api/features
Controls how closely the model follows the input image
Controls the amount of motion in the generated video

Generated cURL Command:

curl -X POST "https://api.1min.ai/api/features" \
  -H "API-KEY: <your-api-key>" \
  -H "Content-Type: application/json" \
  -d '{
  "type": "IMAGE_TO_VIDEO",
  "model": "stable-video-diffusion",
  "conversationId": "IMAGE_TO_VIDEO",
  "promptObject": {
    "image": "<base64-encoded-image-or-url>",
    "seed": 12345,
    "cfg_scale": 2.5,
    "motion_bucket_id": 127
  }
}'

Parameters​

Required Parameters​

  • image (string): The source image for video generation. Can be a base64-encoded image or image URL
  • seed (number): Random seed for reproducible generation
  • cfg_scale (number): Controls how closely the model follows the input image
  • motion_bucket_id (number): Controls the amount of motion in the generated video

Parameter Details​

image​

  • Type: String
  • Required: Yes
  • Description: The source image to animate. Accepts base64-encoded images or image URLs.

seed​

  • Type: Number
  • Required: Yes
  • Range: 0 - 4,294,967,295
  • Description: Random seed for reproducible generation. Use the same seed with identical parameters to get consistent results.

cfg_scale​

  • Type: Number
  • Required: Yes
  • Range: 0.0 - 10.0
  • Default: 2.5
  • Description: CFG (Classifier-Free Guidance) scale controls how closely the model follows the input image. Higher values result in outputs that are more faithful to the input image but may be less creative.

motion_bucket_id​

  • Type: Number
  • Required: Yes
  • Range: 1 - 255
  • Default: 127
  • Description: Controls the amount of motion in the generated video. Lower values produce subtle motion, while higher values create more dramatic movement.

Advanced Parameter Guidelines​

CFG Scale Recommendations​

  • 0.0-1.0: Very loose interpretation, highly creative but may deviate significantly
  • 1.0-3.0: Balanced approach with good creativity and adherence
  • 3.0-5.0: Close adherence to input with moderate creativity
  • 5.0-10.0: Very strict adherence, minimal deviation from input

Motion Bucket ID Recommendations​

  • 1-50: Minimal motion, subtle animations
  • 51-100: Light motion, gentle movements
  • 101-150: Moderate motion, noticeable animation
  • 151-200: Strong motion, dynamic movement
  • 201-255: Maximum motion, dramatic animation

Response Format​

{}

Example Request​

curl -X POST https://api.1min.ai/api/features \
  -H "API-KEY: your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "IMAGE_TO_VIDEO",
    "model": "stable-video-diffusion",
    "conversationId": "IMAGE_TO_VIDEO",
    "promptObject": {
      "image": "data:image/jpeg;base64,/9j/4AAQSkZJRgABAQEAYABgAAD...",
      "seed": 42,
      "cfg_scale": 2.5,
      "motion_bucket_id": 127
    }
  }'

Features​

  • Reproducible Results: Use seeds for consistent generation
  • Fine-grained Control: Advanced parameters for precise customization
  • Professional Quality: StabilityAI's proven diffusion technology
  • Flexible Motion Control: Wide range of motion intensity options
  • No Prompt Required: Works directly with images without text prompts
  • Research-Grade Model: Based on cutting-edge video diffusion research

Best Practices​

Parameter Tuning​

  • Start with default values and adjust incrementally
  • Use consistent seeds for A/B testing different parameter combinations
  • Test motion bucket values in increments of 25-50 for noticeable differences
  • Keep CFG scale between 1.5-4.0 for most use cases

Image Preparation​

  • Use high-resolution images for best results
  • Ensure good lighting and contrast in source images
  • Images with clear subjects tend to produce better motion
  • Consider the intended motion when selecting motion bucket values

Use Cases​

  • Research and Development: Reproducible results for academic work
  • Professional Video Production: High-quality animation for commercial use
  • Creative Experimentation: Fine-tuned control for artistic projects
  • Prototype Development: Consistent results for iterative design
  • Technical Applications: Precise control for specific motion requirements