How it works
Returns each video model to pass to generate_video, with its supported durations, resolutions, aspect ratios, reference limits and whether it makes sound. Use it to avoid combinations a model can't do.
Inputs
No inputs - call with an empty arguments object.
Returns
{ models: [{ model, label, native_audio, modes }] , edit_video, motion_control }Examples
Find a model with sound
The agent picks a model with native audio for a talking ad.
{}