Burning Bridges AI Video Prompts
These prompts are written by us from the structure the published guides describe. Treat them as starting points and adjust the details to your photo and tool. None of them contain song lyrics; add the official sound when you post.
- Name the cup hand in every prompt, or it switches sides.
- For swaps, say what to keep unchanged as clearly as what to replace.
- Keep the gesture as a hand over the eyes, never over the mouth.
@Video1and@Image1are the reference tags used by swap tools; use your tool's own syntax if it differs.
1. Character swap on the clip
For Genjutsu Object Swap or a reference-to-video model. One reference clip, one photo.
Replace the lead man waving at the camera in @Video1 with the person in @Image1. Match his pose, wave, cup, cup hand and position in the frame exactly. Keep everything else from @Video1 unchanged: the dining room, chandeliers, warm lighting, the other guests, the handheld camera movement and the cuts. Light the new person with the room's practical lights, with shadows and reflections that match the scene.
2. Two-person swap
Replace the lead man waving at the camera in @Video1 with the person in @Image1, matching his pose, wave, cup and position. Replace the person at frame-right with the person in @Image2, matching their position and movement. Do not change any other guest. Keep the room, lighting, handheld camera movement and cuts from @Video1.
3. One photo, no original clip
For image-to-video. Nothing from the official video is used, so the room and moves resemble the scene.
The person from the reference photo stands in a crowded private dining room at a lively party, crystal chandeliers overhead, warm golden light, dark wood panels, friends around them raising metal cups. They hold a small metal cup in the left hand, point at the crowd with the right hand, then wave at the camera with an open palm. Realistic handheld phone camera, starting wide and pushing in toward their face. Natural skin tones, realistic detail, vertical 9:16, 15 seconds.
4. The hand-over-eyes gesture
Add this beat to prompt 3, or use it alone for a shorter clip.
Midway through, the person raises the right hand flat above their eyebrows like a visor and slowly scans the room from left to right as if looking for someone, cup still in the left hand beside their face. The hand stays above the eyes and never covers the mouth. Then they lower the hand and grin at the camera.
5. Pet version
The pet from the reference photo stands upright on its hind legs at a lively chandelier-lit dinner party, holding a small metal cup in its left front paw. It waves at the camera with the right paw, then lifts that paw above its eyes and looks around the room. Keep the same species, fur colour and markings as the photo. Playful handheld phone camera, warm golden light, guests cheering in the background, vertical 9:16.
Pet results vary more than human ones, because video models are trained mostly on people. Check paws, balance and face consistency in each take.
6. Animated or virtual character
Replace the lead man in @Video1 with the character in @Image1, keeping the character's own proportions, silhouette, style and costume. Transfer his wave, cup raise and position onto the character's anatomy. Render the character as a 3D-animated film character lit by the room's practical lights, with shadows and reflections consistent with the scene. Keep the room, the other people and the camera movement unchanged.
A 3D film look holds up better in this room than flat cartoon art, because the lighting can match.
Negative prompt
no duplicate subjects, no new scenes, no extra fingers, no extra limbs, no face changes on background guests, no cup switching hands, no on-screen text, no captions, no watermark, no logo
Timing the moves: a beat list
If your tool accepts long prompts, list the moves with timestamps. The format below follows the one Starrd's guide describes; set the times from your own trimmed clip.
0–3s: cup raised toward the camera, smiling 3–6s: points at the crowd with the right hand 6–9s: open-palm wave at the camera 9–11s: right hand flat over the eyes, scanning the room, cup beside the face 11–15s: lowers the hand, laughs, camera pushes in
FAQ
Do I need a prompt at all?
Not for one-photo templates; they have the routine built in. You need one for character swaps and for image-to-video.
Why does the cup change hands?
The model was not told which hand holds it. State the cup hand in the still prompt and again in the video prompt.
Can a prompt alone recreate the scene?
Only roughly. Without a reference clip the model may invent movements or miss the signature gestures. A reference clip gives a closer match.
Should the prompt include the lyrics?
No. Most video models do not produce the song anyway. Add the official sound inside TikTok, Instagram or YouTube when you post.
Which aspect ratio should I ask for?
Vertical 9:16 for TikTok, Reels and Shorts. Swaps of the original clip come out in the clip's own landscape framing, so crop afterwards.
Unofficial, independent guide. Not affiliated with, endorsed by or connected to Drake, OVO Sound, Republic Records or any AI tool mentioned here. We do not host the music video, the song or any copyrighted footage.