Monet Vision Tutorial: Compare Models Without Wasting Credits

1. Define the outcome

Write audience, channel, duration, aspect ratio, subject, action, camera, audio, required details, and prohibited changes.

Collect approved images, product geometry, logos, scripts, faces, voices, and rights. Mark references that must remain unchanged.

Pick two or three current models based on the shot: native audio, identity, motion control, editing, duration, resolution, or draft speed.

Keep prompt and references constant. Start at the lowest useful setting to compare composition and instruction following.

Choose the strongest result, record the model and settings, then refine one variable at a time.

Use lip sync, video swap, motion control, extender, long-take, or conversational editing only where the shot needs them.

Check anatomy, object permanence, identity, product truth, camera intent, flicker, dialogue, ambience, and synchronization.

Record credits, retries, usable seconds, editing time, and rights checks. Choose the workflow with the best approved-output economics.