How to turn an image into a video with Flow AI Video

To turn an image into a video, open Image to Video AI, upload the photo, and describe what should move. Select a model and supported settings, check the credit estimate, and generate. Create retains your prompt and image references after submission.

Should the image go in Start or End?

Put an opening reference in Start and a closing reference in End. You can prepare an end-only reference; it should remain in End rather than becoming a start image. If you upload both, check the previews before generating, regardless of the upload order.

Choose a clear photo with a recognizable subject and room for the motion you want. For a product, readable packaging and a simple background make the result easier to compare with the original.

How do I describe motion?

Describe the change over time rather than repeating everything visible in the photo. Separate the movement of the subject from the camera movement, and state what should stay still.

Keep the product upright and stationary. Slowly move the camera closer while a soft highlight crosses the surface. Preserve the simple background and avoid extra objects.

Reference images guide the result but do not guarantee exact labels, faces, or object shapes. Review those details throughout the clip.

What happens when I generate?

The selected model determines the available formats, durations, and resolutions. Review credits before submitting. With sufficient credits and a successful submission, the task opens in Create with the same prompt and reference positions.

Watch the completed clip and compare it with the original photo. Look for drifting edges, changing text, flicker, or unintended movement. Download a usable take, or simplify the motion brief before trying again.

Can I start from a video instead?

Yes, use Reference to Video for existing footage. It has a video-reference input; uploading a clip is different from choosing a still start or end frame.

Animate your image or plan a product video.