Paste-ready Veo prompt for the Ramalama trend

The prompt

Image-to-video, vertical 9:16, a single unbroken shot. If the tool limits the length, keep every beat in the same order and shorten each one.
Subject: reference image 1 is the STAR, whose face, hair and skin tone must match the image exactly. A pet, toy or drawing comes to life and walks upright.
Outfit: witch, meaning a tall pointed black hat, a floor-length black dress with a torn hem, stockings striped black and purple, and pointed ankle boots. Leave the face uncovered.
Setting: a calm suburban street late at night, dark road, warm orange street lamps, hedges and trees, no readable writing anywhere.
Camera: a phone held close to the ground, drifting backward ahead of the STAR as they approach. Full frame, no fisheye, no cuts.
Opening (up to about seven and a half seconds): only legs, boots and the dress hem are visible, a bouncy runway stride with a footfall on every beat. No face yet.
Lift (less than a second): on a heavy beat the camera snaps from the feet up to the face, showing the full costume at once.
Strut (about two seconds): the STAR keeps coming, arms swinging, then frames the face with both hands and smiles silently.
Finale (until the end): the STAR leans into the lens for a grinning, laughing close-up and stays there.
Audio: spare electro-pop instrumental, a stomping drum on each step, soft footsteps, no singing, no lyrics, no whispers, no speech.
Finish: sharp, photorealistic, natural colour. No captions, no logos, no watermark.

Using it well

• Attach your photo as the reference image first, and paste the text afterwards.
• You can change the outfit and setting lines. Leave the timed lines alone so the lift stays on the beat.
• Veo may cap clips below 12 seconds, so keep the beats in order and make each one briefer.