Put yourself in the film, speaking. The three-step recipe behind the dragon film: your face from a photo, your real cloned voice driving the lip sync. This is the demo that makes people watch again.

STEP 1 — THE FACE. Save a clean chest-up portrait as a character reference on the platform, then generate a new studio portrait FROM that reference with an image model: "Photorealistic studio portrait of [your character reference], chest-up, facing the camera directly, neutral gray background, soft even lighting, natural confident slight smile, ultra-detailed skin texture." Use THIS generated portrait as your video reference — platform-generated images pass the identity check that blocks raw uploads.

STEP 2 — THE VOICE. Generate each spoken line as text-to-speech in your cloned voice, one line per file, each under four seconds. Attach them to the video generation as audio references along with the portrait.

STEP 3 — THE PROMPT. "A cinematic 30-second [genre] short film with spoken dialogue, starring the man from the reference image with his exact same recognizable face throughout, wearing [outfit]. AUDIO MIX: [score brief that ducks under dialogue]. SCENE ONE: [action]. He speaks, lip-synced exactly to the first audio reference in his real voice: '[YOUR LINE].' SCENE TWO: [action, second line at its beat]. Sound design: [effects list]."

Customize: the line that ties it together is "lip-synced exactly to the audio reference in his real voice" — say it in every spoken moment. One or two short lines per 30-second clip; give the rest of the runtime to the world.