How to Make the Hop Hop Hop AI Video
The joke is the switch: you point toward someone hopping behind you, the shot cuts, and suddenly you are the one hopping. This guide breaks down the night parking lot reference, explains the photo roles and shows how to describe the swap without letting identities merge.
Understand the dance before making it
Hop Hop Hop is also called the Gabi trend. The twin version appears in @gabilagaba's October 18 post, followed by cosplay, emo-outfit and animated versions. The dates identify the linked TikTok posts; they do not establish the first-ever performance.
Our motion reference is @pouzato7's animated parking-lot clip. It has two staggered subjects, three hops, a low bounce and a hard cut. The footage below is a credited third-party reference, not a result made by this site.
Read the reference second by second
The downloaded reference contains 232 frames at 30 fps: about 7.73 seconds, in a 576 × 1024 portrait frame. The template asks the model for 8 seconds. These times describe the source, not a frame-accurate guarantee for a generation.
| Time | Foreground | Background / edit |
|---|---|---|
| 0–1.3 s | A is close to the phone on the right | B is back-left, fully visible |
| 1.3–4.4 s | A grins and points back | B hops three times, then bounces low |
| Around 4.4 s | B replaces A instantly | One hard cut; same camera position |
| 4.4–7.7 s | B gestures with a hand above the eyes | A hops with arms near the head |
The swap is an edit. If the subjects walk past each other, you have lost the defining beat. Ask for one hard cut and no walking transition.
Choose the cast and prepare photos
- Two friends: upload one photo in each slot. Photo 1 points first; photo 2 hops first. They exchange roles after the cut.
- Twins: use only photo 1. The prompt makes two copies of that identity and outfit. Leave slot 2 empty.
- You and a pet: put your photo in slot 1 and one full-body dog or cat photo in slot 2. The pet keeps its species, coat and collar; upright hopping is a stylized effect.
- Framing: a sharp, evenly lit face helps likeness; knees-up or full-body framing helps legs and clothing. Avoid group photos and cropped paws.
Method 1: use the Hop Hop Hop template
Open the maker
Visit the Hop Hop Hop AI generator and choose the cast mode.
Upload the photos
Photo 1 is the first foreground subject. Add a friend or pet in slot 2, or leave it empty for twins. An empty second slot defaults to twins in any mode.
Choose soundtrack and quality
The reference soundtrack is optional. Select No song if you plan to use a sound from the posting app. Check the maker's credit price for the selected quality.
Generate and review
Check the foreground face before and after the cut, count the background subjects and watch the feet and paws. Generation timing can vary.
Download
Save the vertical MP4. Use the platform's AI label when posting and credit the inspiration where appropriate.
Prompt breakdown: identity, staging, action, sound
The following is a reusable prompt recipe. Attach your own photos as identity references and a motion clip if the tool supports it. This describes the observed blocking and adds identity and anatomy constraints for a new generation; it does not recover the creator's private prompt.
| Prompt layer | What to specify | Problem it avoids |
|---|---|---|
| Identity | A = photo 1, B = photo 2; one photo means two identical copies | Faces switching or merging |
| Scene | Night parking lot, bright overhead lamp, white lines, distant trees | A different environment after the cut |
| Camera | Fixed portrait phone; A close-right, B full-body back-left | A moving camera or hidden feet |
| Action | Three hops, point back, one hard cut around 4.4 s, reverse roles | Walking swaps and dissolves |
| Continuity | Same faces, outfits and pet markings across both shots | A new person or species after the cut |
| Sound | Quiet night ambience and soft landing sounds; add music separately | Invented lyrics or an approximate song |
English visual prompt: Create an 8-second vertical phone video using the attached photos only for identity. A from photo 1 stands close-right, seen waist-up; B from photo 2 stands farther back-left, head to feet visible. Keep their own clothes and faces. In a night parking lot under one white streetlamp, B hops three times while A smiles and points back. Around 4.4 seconds make exactly one hard cut: B is now close-right and A is back-left. B gestures above the eyes while A hops with raised arms. Fixed camera and identical background in both shots. No walking transition, face morph, extra subject, captions or logos.
English audio prompt: Quiet outdoor night air and soft shoe or paw landings synchronized with the hops. No dialogue, voiceover, singing or recognizable melody. Add the selected soundtrack separately after rendering.
- Twin variation: replace the B identity line with ‘B is an identical copy of A from photo 1, in the same clothes; exactly two copies’. Use no second photo.
- Pet variation: specify ‘B is the animal in photo 2, preserving species, muzzle, coat pattern and collar. Stylized hind-leg hops; a forepaw gesture after the cut. Never a human body or human hand’. Results can vary.
- What is creative guidance: identity preservation, anatomy constraints and the quiet generated ambience are production instructions added for this template. They are not claims about the source creator's prompt or recording method.
Handle the sound separately
The trend uses the Russian hook commonly called ‘Её парень’ (‘Her Boyfriend’) and attributes the TikTok sound to @venadeaakira. We have not verified a complete lyric translation or an original release date. The downloaded clip has an AAC stereo soundtrack. The maker can place its reference audio under the result, or you can switch it off and select the relevant sound in the posting app.
A music title in a prompt will not reproduce a specific recording or reliably align the swap. For a manual edit, place the cut against the audio waveform and check playback.
Method 2: film and edit it yourself
Lock the phone
Keep the phone fixed in portrait orientation and mark the two standing positions. Leave space above the back subject's head and below their feet.
Film the first arrangement
A points toward B while B does three hops. Keep the foreground reaction simple.
Swap positions and film again
Use the same camera and exposure. B reacts near the phone while A hops behind.
Cut to the beat
Join the two takes at the beat change using a hard cut. Do not add a dissolve. Trim the ending and export vertically.
To make twins manually, film the same person twice with the camera locked and use a mask or split composite. Keep the two performances on different sides of the frame so their bodies do not overlap.
Common problems and fixes
- The characters walk through each other: explicitly request an instantaneous hard-cut exchange, not a continuous movement.
- A face changes after the cut: repeat both identity assignments and keep outfits unchanged across both shots.
- Feet leave the frame: place the hopping subject farther back and ask for full-body framing.
- The pet becomes human or too large: specify species, paws and realistic relative scale; use a full-body photo. A new generation may still be needed.
- The source cartoon costume appears: state that the source supplies motion only and the clothes come from the uploaded photos. The current reference is RGB footage, so unwanted appearance carryover remains possible.
Watch five credited trend references
These are third-party posts, not maker outputs. Dates and rounded play counts come from an October 30, 2026 snapshot. The remake example uses the supplied sample video and carries an AIVideoTrend watermark.
Twin dance in a poster-filled bedroom; the front dancer points and the back dancer hops.
Animated characters do the swap in a parking lot at night. This is the motion reference.
Two friends in dark emo outfits repeat the point, hops and swap indoors.
MIKU VS MIKUO: the same dance with a cosplay pair.
My Little Pony cosplayers take turns hopping outdoors at night.
常見問題
How many photos do I need?
One for twins; two for distinct friends or a person and a pet. Photo 1 points first; photo 2 hops first.
What is the most important part of the prompt?
Assign two identities, keep the phone fixed and ask for one hard cut around 4.4 seconds that exchanges their positions.
Does this exactly reproduce the original?
No. The reference guides the motion, but likeness, choreography and cut timing can vary. The requested output is 8 seconds; the downloaded source is about 7.73 seconds.
What does the remake tab show?
It plays the supplied sample video after the lower-edge watermark was cropped away and the AIVideoTrend watermark was added.
