Skip to main content
Image to Video converts static images into dynamic video sequences. You provide one or two key frames — a start image and optionally an end image — and the AI generates the motion and intermediate frames that connect them into a smooth, continuous clip. This mode is ideal when you already have a specific visual in mind and want to bring it to life, rather than generating imagery from scratch via a text prompt.

​
How to use Image to Video

1

Open Video mode

Navigate to Video mode from the left sidebar to open the Video Studio.
2

Select the Image to Video mode

Click Add image below the prompt field. This opens the input modes tray. Select Image-to-Video to activate the frame upload interface.
3

Upload your start frame

Upload the image you want the video to begin from. This image defines the opening scene — the AI uses it as a precise anchor and builds outward from this frame.For best results:
  • Use a high-resolution image (minimum 300px on the shortest side)
  • Prefer JPG, JPEG, PNG, or WEBP formats
  • Keep the file under 10 MB
  • Choose images with clear subjects and unambiguous composition
4

Upload an end frame (optional)

If you want to control where the video ends, upload a second image as the end frame. The model will generate the transition between your two images, interpolating motion, lighting, and perspective changes to create a coherent sequence.
When using both start and end frames, choose images that share some visual continuity — similar subjects, environments, or colour palettes — to help the model produce a plausible and visually coherent transition.
5

Configure settings

Select the aspect ratio, duration, and resolution that match your output requirements. The available options depend on the model you choose.Not all models support end frames. Models that do are marked with Start/End frame in the credit consumption table.
6

Generate the video

Click Generate. The AI produces a video that transitions smoothly from your start frame, and to your end frame if provided. Generation typically completes within 30–60 seconds.

​
Example use cases

Use a daytime landscape as the start frame and a night version of the same scene as the end frame. The model generates a natural time-lapse-style transition, shifting light, shadows, and atmosphere across the clip.
Upload a still of a character or vehicle and use a text prompt alongside the image to describe the intended motion — for example, “a racing car accelerating from a standing start.” The AI animates the subject based on both the visual and descriptive inputs.
Use two images with different lighting or compositional states to create an abstract visual journey. Gradual shifts in colour, background detail, or perspective can produce compelling artistic animations with minimal effort.

​
Supported input formats

InputRequirements
FormatJPG, JPEG, PNG, WEBP
Minimum resolution300px (shortest side)
Maximum file size10 MB per image
Maximum imagesUp to 4 images total per generation
Image to Video uses the same video models as Text to Video. Each model has different support for start-only vs start-and-end frames. Check the tooltip in Video mode or the credit consumption page to confirm what a specific model supports before uploading.

​
What to do next

Reference to Video

Use multiple reference images to guide the style and content of a new video, rather than defining start and end frames.

Extend Video

Add 5 more seconds to the end of your generated clip.

Video Effects

Apply cinematic VFX effects to your animated output.

Video Credits

See how credits are calculated for Image to Video generations.