AI Video with Runway
Runway Gen-3 generates video from text prompts or still images. The results are increasingly cinematic, smooth motion, realistic physics, and controllable camera movement. This is the frontier of creative AI.
Start from an image, not text. Generate a still in Midjourney, then animate it in Runway. This gives you much more control over the starting composition.
Camera motion prompts: 'slow push in', 'tracking shot left to right', 'drone shot rising above'. Runway understands cinematic language.
Keep clips short (4-10 seconds) and edit them together. AI video works best in short bursts stitched into a sequence.
Motion brush lets you define which parts of the image should move and which should stay still. Essential for controlled, natural-looking results.
Sign up for Runway (runwayml.com). Free tier available, paid for full features
Choose your input: text prompt, uploaded image, or image + text
For image-to-video: upload your source image. Add a motion description
Set duration (4-10 seconds for best quality)
Use Motion Brush to control which areas move: paint over water to make it flow, paint over hair to add wind
Generate and review. Iterate on the motion description
Download and edit clips together in a video editor (DaVinci Resolve is free and professional-grade)
Combine with AI-generated images for a full AI-produced video workflow