An On The Radar AI prompt is a written scene description that tells an image-to-video model to put the person or pet from your photo into a green freestyle booth and make them rap for about fourteen seconds. Below you will find four prompts to copy (a one-person prompt for other tools, our two-person video prompt, a pet prompt and the scene-image starter), a line-by-line breakdown of what each part controls, and how our maker combines a preset with optional Director’s Notes.
Last reviewed Oct 19, 2026 · On The Radar AI editorial team
Rather not write anything? The maker runs a fixed scene prompt on your photos.
What a Good On The Radar AI Prompt Has to Do
The booth is easy to describe. The hard part is discipline: your face must stay yours, the camera must stay still, two performers must not blend, and the soundtrack must not drift toward a song someone else owns. Each On The Radar AI prompt below handles those jobs in one order: format, room, framing, motion, mouth, identity, sound, then a short never-show list.
The wording is plain English for any image-to-video tool that accepts one or two reference images plus text. Our maker sends its own version to ByteDance Seedance 2.0 and tags uploads as @Image 1 and @Image 2; other tools use their own tags, so replace "the first reference image" with whatever your tool documents. We tuned the wording on Seedance 2.0 only, so treat other models as a starting point.
One-Person Booth Prompt (For Other Tools)
Our maker always takes two photos. If you want a one-person booth clip in another image-to-video tool, use this with a single sharp, front-facing photo of yourself or of an adult who said yes. A waist-up or chest-up shot gives the model the most to work with.
Vertical 9:16 video, 14 seconds, photorealistic.
Performer: the person in the first reference image is the only performer. They stand behind a studio microphone with a round pop filter, centered, framed from the waist up.
Room: a small freestyle booth. The wall behind them is flat and painted deep green, evenly lit. A plain glowing ring light hangs on that wall: an empty circle of light with no words, letters, numbers or logo in or around it.
Camera: a static medium shot for the first 3 seconds, then a slow, smooth push toward the microphone until the end. One continuous take, no cuts, no handheld shake.
Action: the performer bobs their head to the beat and raps into the microphone with loose, confident hand gestures. Their mouth opens and closes in rhythm with the vocal.
Identity: same face, hairstyle, skin tone, age and outfit as the reference image in every frame. Nobody else enters the shot.
Sound: an original hip-hop beat with punchy drums and an English rap verse. It must not resemble any released song, melody or lyric. No crowd noise.
Never show: captions, subtitles, on-screen lyrics, watermarks, brand names or logos.
If your tool makes silent video, delete the Sound paragraph and add audio later in your editing or posting app.
Two-Person Booth Prompt With Locked Left and Right Sides
Our maker's workflow uses a composed image as @Image1 and the template performance as @Video1. The video prompt binds the left and right subjects to their corresponding performers.
@Image1 shows the TWO characters. @Video1 is the performance to copy.
The character on the LEFT of @Image1 copies the LEFT performer's moves. The character on the RIGHT of @Image1 copies the RIGHT performer's moves.
The subjects in @Image1 do the same exact moves as @Video1 for the whole video.
The entire clip is the two characters from @Image1. Every frame from the first to the last shows only them. Preserve each one's exact identity: face, hair, skin tone, fur, markings and clothing, exactly as in @Image1. Do not blend them, do not swap their clothing, do not add or remove characters. Animals stand upright on their hind legs.
Use the composed two-person studio image, not the two raw photos, for @Image1. Supply the performance clip as @Video1, render 14 seconds without generated audio, then add the aligned soundtrack.
Pet Booth Prompt for a Dog, Cat or Any Animal
For a one-animal clip in another tool, animals need two extra rules: stay the same animal, and rap in a playful way instead of a human way. A clear, upright photo where the face and both eyes show works best.
Vertical 9:16 video, 14 seconds, photorealistic with a playful tone.
Performer: the animal in the first reference image sits upright on a stool behind a studio microphone, centered in frame.
Animal rules: it remains exactly the same animal, with the same species, breed, coloring, pattern and markings as the reference image. It does not turn into a person and wears nothing it is not already wearing in the photo.
Room: a small booth with a flat deep green wall, soft overhead light and one plain glowing ring light with no text or logo.
Camera: static medium shot, then a slow push toward the microphone near the end. One take, no cuts.
Action: the animal bobs its head and ears to the beat and "raps" in a comic, cartoonish way, its jaw opening and closing roughly in time with the vocal.
Sound: an original upbeat hip-hop instrumental with an English rap vocal; no recognizable song.
Never show: captions, watermarks or logos.
For a pet and owner together, use the two-person prompt instead: it already tells the model that animals stand upright on their hind legs. In our maker, a pet simply fills one of the two photo slots.
Two-Person Scene-Image Prompt
Some people get steadier results in two passes: first an image model edits the photo into a booth portrait, then the video model animates that frame. It costs two generations, but you approve the look before paying for motion.
A photorealistic VERTICAL 9:16 studio photograph of the TWO subjects from the reference photos performing together in a radio studio.
LEFT subject: the first subject from the reference photos, standing TALL and FULLY UPRIGHT — body vertical, balanced and stable, NOT sitting, NOT crouching. If the subject is an animal it stands on BOTH hind legs in a confident human-like posture. Arms or front paws raised and clearly separated from the body and from each other at about chest height, as if gesturing while rapping.
RIGHT subject: the second subject from the reference photos, standing TALL and FULLY UPRIGHT on both feet, arms down and clearly separated from the torso, relaxed and confident.
PRESERVE the exact identity of each subject — face, hair, skin tone, breed, fur, markings and clothing — exactly as in their reference photo. Only the pose and the setting change. Do NOT blend the two subjects together and do NOT swap their clothing.
They stand side by side with a clear gap of empty floor between them, neither overlapping nor touching, each forming ONE clean unobstructed silhouette; the microphone stand rises through the gap between them and does not cover either face. BOTH are shown FULL BODY from head to feet, feet flat on the floor, with clear headroom above both heads and visible floor below both. The pair is centered and fills MOST OF THE FRAME HEIGHT.
SETTING: a radio studio. Behind them a wall washed in deep green light carries a single large glowing white neon sign reading "ON THE RADAR" inside a circular badge, with tall vertical green neon tube lights to either side and dark patterned acoustic panels. A low dark green sofa sits against the wall behind them and a shelving unit stands at the right edge. ONE black condenser microphone on a slim boom stand stands in the FOREGROUND, closer to the camera than the subjects, centered on the gap between them: the microphone head is at mouth height, midway between the two faces, and the stand runs straight down through the middle of the frame to its round base on the floor at the bottom edge, in front of the floor the subjects stand on. The microphone is between the subjects and the camera, never behind them and never level with them. Smooth grey studio floor. Moody green-tinted studio lighting.
Photorealistic, sharp focus. Vertical 9:16 portrait orientation. No text or captions other than the neon "ON THE RADAR" sign itself, no logos, no watermarks, no other people or animals.
Supply the two original photos in left/right order. The resulting scene image becomes the single image reference for the video step.
Prompt Anatomy: What Each Part Controls
A one-person prompt for another tool can be a single direct video prompt. Our two-person workflow separates scene composition from performance: describe the image first, then map its left and right subjects to the reference video.
Part
What to write
What happens if you skip it
Format
Vertical 9:16, about 14 seconds, photorealistic
Landscape output you have to crop, or a clip that runs long and loses the face
Booth set
One person: plain green booth. Two people: the detailed green studio in the scene-image prompt.
The model invents a busy studio or paints a show's sign onto the wall
Visible text
Specify required set details and exclude added captions or watermarks.
Garbled lettering inside the ring, fake captions or a made-up watermark
Framing and camera
One person: a gentle push-in. Two people: follow the reference performance.
Drifting handheld moves, sudden cuts or a camera that circles the subject
Motion
Two people: left subject copies left performer; right subject copies right performer throughout.
Stiff mannequins or random dancing that ignores the beat
Mouth
Mouth moves with the vocal; with two people, mouths follow the reference performance
Both mouths flap at once, or nobody seems to be rapping
Identity
Same face, hair, skin tone, age and outfit as the photo; no extra people
Faces drift toward a generic look, or a stranger walks into frame
Audio
One person: generated sound, if your tool has it. Two people: render silently and add the aligned template track.
Silence, muddy noise or something that sounds borrowed from a real record
Mouth movement is an approximation of rapping, not true lyric sync. No model we know of reliably matches your own words or voice from a text prompt.
Quick Fixes When a Booth Clip Comes Out Wrong
Most problems trace back to the photo or to one sentence missing from the On The Radar AI prompt. Try these before rewriting from scratch.
The face drifts. Use a sharper, well-lit, front-facing photo of one subject with nothing covering the eyes, and keep the identity paragraph intact.
Left and right swap. Upload the left performer first, add one clothing detail per person and keep the left/right mapping lines intact.
Letters appear in the ring light. Keep the empty-circle sentence and never type a show, channel or artist name.
Hands or teeth look odd. Ask for relaxed gestures and pick a photo where the hands are not blurred or cropped.
The pet looks too human. Keep the animal rules and the stool; a rapping dog will always look a little cartoonish.
The beat sounds familiar. Regenerate, then add the music you actually want inside TikTok or Instagram.
Why Our Maker Needs No On The Radar AI Prompt
Upload two photos, one per performer. The maker prepares a scene image with GPT Image 2, animates it with Seedance 2.0 and the fixed performance reference, then adds the template soundtrack. Your optional Director’s Notes customize the scene image.
You never type or edit a prompt in the maker. Its scene-image and video prompts come from the captured template configuration and run on the server, and the video performance stays tied to its reference clip. The copy-ready prompts above follow the same structure for other video models.
Ground Rules Before You Paste Any Prompt
Only animate yourself, adults who have agreed to appear, or animals you are allowed to photograph. Leave out minors, celebrities and the artists from the original performance, and do not write a real show, channel or rapper into the prompt: it invites logos and lookalikes you have no right to publish. Our privacy policy explains how uploaded photos are handled in our maker, and the terms cover what you may post.
When you share the result, keep your platform's AI label switched on and describe it as an AI clip, not as an appearance on the real series. On The Radar AI is an independent site with no tie to On The Radar, its founder, Fetty Wap or Monty; the trend explainer covers where the format came from.
06 · FAQ
Questions People Ask
Not inside the maker: its prompts run on the server, so there is nothing to paste or tune beyond the optional Director’s Notes for the scene image. The two-person prompt on this page follows the same structure if you want to try it in another video model.
Any image-to-video tool that accepts a reference photo plus text can try them. We wrote and tested the wording on Seedance 2.0, the model our maker uses, so on other models expect to adjust the image tags and, if the model is silent, drop the sound section.
Use optional Director’s Notes to customize the scene image. The performance prompt stays tied to the template’s reference video and left/right mapping. Model, duration and resolution settings are controlled by the maker.
Our maker adds the 679 template soundtrack after the video renders. A text prompt alone does not supply the original recording.
Models track position loosely. Upload the person who belongs on the left first, state which subject is left and which is right, and add a clothing detail for each. In our maker, the Swap sides control flips the order before you render.
Our scene-image prompt includes the ON THE RADAR text for the studio sign. Your uploaded photos define the performers. The video prompt binds the composed image to the performance reference.
Only up to a point. What matters is covering the eight parts in the anatomy table once and clearly. Repeating adjectives or stacking style words tends to blur the instructions rather than sharpen the result.
Next stop
Where to Go From Here
Each page below tackles one question about the booth clip, and every one ends at the same maker.
Skip the On The Radar AI prompt entirely: bring two photos, choose 480p or 720p and the booth scene is applied for you, with the cost shown before rendering.
