Quick answer
Starrd's Stop Humiliating Yourself template puts you in the Rush Hour 3 squad-room scene. Upload one to three photos: photo 1 is the yeller, who grabs the suspect by the collar, shouts mock French back at them and ends on 'Stop humiliating yourself!'; photo 2 is the suspect, who answers in fake French; photo 3 is the partner, who can only say 'How can I tell him? I don't speak French.' Any role without a photo gets an AI-cast stranger, and if every photo is a pet the strangers are animals too, all in tiny black suits. Everyone else keeps the outfit from their photo, every line is lip-synced to the original audio, and all ten cuts land on the same frames as the film. 19 seconds, widescreen 16:9, Seedance 2.5 Flash, 480p, 200 credits (about $10).
The "Stop humiliating yourself!" Rush Hour interrogation, with you doing the yelling, or your cats. Here it is with three cat photos:
Stop Humiliating Yourself
1-3 photos. Your suspect babbles fake French and you lose it. 19 seconds, people or pets.
What You Get
- The grab. You hold the suspect by the collar and shout their fake French right back at them.
- The suspect. Smug, theatrical, answering only in gibberish French.
- The partner. A half-second glance from the counter, then a deadpan close-up: "How can I tell him? I don't speak French."
- The punchline. A close-up of you exploding: "Stop humiliating yourself!"
- The squad room. Cubicles, filing cabinets, fluorescent lights, a 2000s action-comedy look.
- Lip sync, your own outfits, the original sound, 19 seconds, widescreen.
What the Video Actually Is
The scene is from Rush Hour 3 (2007). The detective and his partner have a suspect in the squad room, and the suspect decides the best defence is to speak only French. The detective answers by yelling the gibberish back, his partner is no help, and the scene ends on the line everyone quotes.
It keeps coming back on TikTok through "Best of Rush Hour" clip series. This version is built from a November 21, 2026 repost.
Why it suits AI: the joke is the faces and the timing. A cat in a suit yelling "Stop humiliating yourself!" at a kitten in a beanie is the version people send to friends.
The 19 Seconds, Shot by Shot
| Time | What happens |
|---|---|
| 0:00 | Medium two-shot: the yeller grips the suspect's collar and shouts mock French |
| 2.17s | Half a second on the partner at the counter, glancing back |
| 2.67s | Close-up: the yeller yells more mock French, then freezes wide-eyed |
| 3.57s | Close-up: the suspect answers in bursts of fake French |
| 5.30s | Wide: "What the hell is that?" ... "Tell him to stop playing around and talk right." |
| 10.43s | Extreme close-up of the partner: "How can I tell him? I don't speak French." |
| 12.90s | Whip-pan, then a wide shot of the yeller shaking the suspect: "Talk right, man!" |
| 15.30s | Close-up: the suspect jabbers more French, defiant |
| 16.90s | Close-up: "Stop humiliating yourself!" and a held glare to 19s |
Your Look, Your Roles
- Photo order sets the roles. Photo 1 is the yeller, photo 2 the suspect, photo 3 the partner.
- Your own outfits. Whatever each person wears in their photo, they wear in the squad room.
- Empty roles get strangers. Upload one photo and the suspect and partner are cast for you: people if you're a person, other animals if you're a pet.
- Pets get suits. Every animal wears a tiny black suit, sized for a real pet body.
Put the loudest friend in photo 1: the yeller talks in seven of the ten shots and gets the punchline. A clear, front-facing photo keeps the likeness best.
The Fastest Way
Stop Humiliating Yourself
1-3 photos in the Rush Hour squad room, lip-synced to the original sound.
One to three photos, 200 credits (about $10), no prompt. It renders on Seedance 2.5 Flash at 480p, widescreen 16:9, 19 seconds.
Build It Yourself
1. Crop the clip, then turn it into a soft depth map. The repost is a widescreen film strip inside a vertical video, with captions above and below. Crop the strip, run it through a depth-estimation model (Depth Anything V2 works), then blur everything except the mouths. A soft body keeps the actors' head and shoulder shapes out of your render; sharp lips keep the lip sync.
2. Make a character sheet. The empty squad room across the top and a full turnaround of each person below, with the room softly out of focus behind every panel.
A photorealistic VERTICAL 9:16 CHARACTER REFERENCE SHEET in four bands: the empty police squad room on top, then a four-angle turnaround of each subject, grey cubicles and filing cabinets softly out of focus behind every panel. No text.
3. Drive it with the depth map and the audio. Feed the sheet, your photos and the depth video, with the dialogue inside it, to a model that copies motion from a reference video (we recommend Seedance 2.5 Flash). Then write what it can't read from a depth map:
- Who talks when: three speakers, so give each one their own mouth windows, measured in frames.
- Who's off screen: the partner's first line and the start of "Talk right, man" are spoken off camera, so every mouth on screen stays shut.
- The shot list: ten shots, each exactly once, with the cut times.
Lay the original audio back on at the end.
Common Mistakes
Trusting the face tracker on the opening. The yeller is in profile and in shadow, and automatic mouth tracking reads him as silent. He's actually shouting the mock French himself. Check the first shots frame by frame.
Muting the wrong voice. Mute only the lines whose speaker is off screen in the audio you feed the model. Mute too much and the mouths stop moving where they should.
Blurring the lips with everything else. A softer driver helps likeness, but blurred lips make the sync drift. Keep a sharp patch on every mouth.
Losing the half-second shot. The partner's glance at 2.17s is so short a soft driver can merge it into the next shot. Zoom the driver in on the partner and say in the prompt that it lasts half a second.
FAQ
How many photos? One to three.
How long is it? 19 seconds, widescreen 16:9.
What does it cost? 200 credits, about $10.
Pets? Yes, in any role, in tiny black suits.
Related Reading
- How to Make the 'Push The G*ddamn Button' AI Video: another Rush Hour scene for up to three
- How to Make the 'Fantasy (Rush Hour)' AI Video: the Rush Hour back-seat sing-along
- How to Make the 'Who Said Something About Oil?' AI Video: this month's other widescreen comedy meme
- Viral AI Video Trends 2026: every trend we've turned into a template this year