Layer generates video the same two ways it generates everything else: describe what you want to the Agent and let it handle the model and settings, or open the generation form yourself for exact control over the model, frames, duration, and aspect ratio.
Generate video with the Agent
In any chat, describe the video you want — "turn this into a short flying loop," "animate this character walking forward" — and the Agent picks a suitable video model, sets the duration and aspect ratio, and generates. If you attach or select an image first, the Agent uses it as the starting point for the animation.
Generate video with the direct form
From inside a session, click Generate video on the bottom toolbar to open the generation form:
Model — pick the video model you want to use.
Prompt — describe the motion, camera movement, and scene.
First Frame / Last Frame — anchor the clip to specific images so you control exactly where the motion starts and ends. See Creating videos with First Frame/Last Frame.
Reference sets — guide the overall look of the video.
Duration and aspect ratio — set the length and shape of the clip; both vary by model.
💡 The same form handles image, video, 3D, audio, and playables — switch modality from the bottom toolbar, then pick a model that fits. See Generate and edit assets directly for the full rundown of the direct generation form.
Turn an existing image into video
You don't need to start from scratch. Select any image on the board, or open it in the lightbox, and choose Turn into video — or just ask the Agent to "turn this into a video." Your image is loaded as the source, and the result lands on the board as a new asset; your original image stays untouched.
What's next
Guide the start and end of a clip → Creating videos with First Frame/Last Frame
Reuse a frame from any video as a reference → Creating consistent videos with Select Video Frame
Pick the right model → Base Models
Keep a consistent look → Using reference sets