Generating Video on Leonardo.Ai

Video is Leonardo.Ai's dedicated motion tool, capable of breathing life into your ideas. Leonardo.Ai offers state of the art video models including Veo 3 and Kling Video 3.0, capable of coherent video that includes sound, allowing you to generate quality videos that include sound effects and audio dialogue.

Whether you're a content creator, a filmmaker creating storyboards, visualizing pre-production assets, or just looking to bring your generations to life, Leonardo.Ai's Video features save valuable time and resources typically spent on video production and animation.

Please note that due to high running costs of some third party models, not all video models are available to free users and there may be usage limits for paid users.

Generating a video from a text prompt

2. Click the side bar to choose a model. You'll find labels that indicate each model's capabilities, including start and end frames, audio and more.

Optional:

3. Enter your prompt in the prompt bar, and Generate. Generated videos live in your Library.

Notes:

Turning generated or uploaded images into a video

To upload an image to be used for video, click the image guidance button in the prompt bar to add your start frame. For more details, read this detailed guide on how to upload images for image and video generation.

Motion Controls & Motion Elements - compatible with Motion 2.0 only.

Motion Controls are camera controls and visual effects that let you control the story in every video.

Motion Elements act like visual effects or filters to help achieve a certain vibe or aesthetic.

Motion 2.0 also offers style options that let you guide the overall aesthetic vibe (e.g stock footage), lighting (e.g lens flare) and color theme (e.g pastel). They are not compatible with Motion Controls or Motion Elements.

Editing videos

2. Enter a descriptive change you want to see - no prompting skills required!

Notes on video editing functions:

Editing is available with some video models such as Kling Video 3.0. However, depending on specific model's capabilities, it may not actually 'edit' your video, but will regenerate it with the new changes you wish to see. It is not the most reliable way of making changes to a scene and maintaining consistency, but is ideal for rapid iterations and experimentation.

If consistency is important, a tip is to edit your images (learn more) to get the exact scenes and variations you want, and use them as start frames for ultimate control.

You may describe how the camera should behave in the prompt including the speed at which it moves, though the effects may not be consistent or 100% effective:

​Dolly in / out - Moves in / out: The camera moves in or out of the scene.

Pull Back: The camera pulls (dollys) back.

Zoom in / out: The camera zooms in or out (This is somewhat different to a dolly)

​

Pan left / right: The camera rotates sideways depending on direction stated. (Note: the model may not always adhere to the direction stated)
​Orbit around: The camera moves around an point.

Tilt: Tilts camera either up or down. (This does not work in most cases)

​​Tracking shot: To use this, it is essential to state what the camera should follow in the shot. (eg, The camera follows the woman in a slow tracking shot)

​​Pull Focus / Focus Racking: It may be possible to shift the focus from one subject to another. To do this, describe how the focus should shift between the subjects in the prompt. It can be described as such: focus pull, the camera shifts focus on a man in the foreground after looking at a woman in the background.

Negative prompts can be inputted for Text to Video by enabling the Negative Prompt toggle found in Advanced Settings. This is useful for shifting the aesthetics and movements away from certain biases.

To use this, simply enter what you do not want to see. Avoid entering negatives such as no or without as this results in double negatives. (eg. No 3D render should be entered as just 3D render)

​