How to turn an AI companion image into a video
Animate an image from your collection
- Open an image in your collection that has the Generate Video action. The action is available for eligible images you own or have saved to your library.
- Select Generate Video to open the image preview and video settings.
- Choose the available duration and resolution. You can leave the motion prompt empty, but a brief instruction usually gives the animation a clearer direction.
- Check the token cost shown in the dialog, then submit. The new clip appears in your collection when processing finishes.
This uses the selected image as the starting frame. It is not a general upload-and-animate tool for arbitrary photos from your device.
Write a motion prompt that fits the image
Describe one action involving the character or scene already visible. If the image shows a person beside a café window, for example, try: “She looks toward the window, then smiles at the camera.” You can also leave the field empty and let the model choose the motion. Results vary, and an instruction that fights the source image may be less predictable.
The dialog offers the current clip lengths and resolutions. A longer or higher-resolution clip can cost more; the displayed price is the one to check before submitting. Video generation requires an active subscription and enough tokens for the selected settings.
Image-to-video or prompt-to-video?
Use this image-to-video workflow when you want to keep a particular starting image and see it before paying for the animation. Use the video generator when you want to describe a new scene in one request: that tool first renders a still frame, then animates it and delivers the clip. If you need a new starting image, make one in the image generator first.
All requests must follow the content rules. A prompt does not guarantee a particular pose, movement, speech or final frame.