Image-to-video gives you much more control over the output by starting with a reference image.

Image-to-Video Workflow

1. Upload a source image (photo, AI-generated image, artwork)
2. Add a text prompt describing the desired motion
3. Set camera movement and duration
4. Generate the video

Why Image-to-Video is Often Better

• Predictable composition — The first frame matches your image exactly
• Style consistency — Maintain a specific visual style throughout
• Creative pipeline — Generate images in Midjourney/FLUX → animate in Runway
• Character consistency — The subject looks exactly how you designed them

Motion Brush

Motion Brush lets you paint motion onto specific areas of an image:

1. Upload or select an image
2. Use the brush tool to select areas
3. For each painted region, set:
   • Direction — Which way should it move?
   • Speed — How fast should the motion be?
   • Ambient — Add subtle environmental motion
4. The rest of the image stays static

Example: A portrait where only the hair moves in the wind, or a landscape where the water flows but the mountains are still.

Multi-Region Motion

Paint different regions with different motions:

• Water flowing left
• Clouds drifting right
• Leaves falling down
• Subject remaining still

This creates sophisticated, controlled animations from a single still image.

Tips for Image-to-Video

• Use high-resolution images (at least 1280x720)
• Simple, clear compositions work better than busy scenes
• Describe motion relative to the image: "the person turns their head to the right"
• Combine with extend feature to create longer sequences