Quick answer
Starrd's I Just Seen A Rainbow template turns one photo into the viral selfie clip: you stare into the phone dead serious insisting how tough you are, whip the camera round to film a huge rainbow over a small-town lawn while you keep talking, a stranger wanders past, and the phone swings back to an extreme close-up of your face, completely won over. One unbroken handheld take with the original audio, every word lip-synced. Contains explicit language. It works with pets too. 1 photo, 13 seconds, vertical 9:16, 480p, Seedance 2.5, 150 credits (about $7.50).
The hardest person you know, undone by a rainbow:
I Just Seen A Rainbow
One photo. Act tough into the selfie camera, film a giant rainbow, then melt. 13 seconds, every word lip-synced.
What You Get
- 13 seconds, one unbroken take, shot like a real phone selfie, whip pans and all.
- A huge, bright rainbow arcing over a wet green lawn, street lamps, power lines and low brick buildings on a grey evening after the rain.
- Your own face and outfit, from the deadpan opening stare to the extreme close-up at the end.
- Every word lip-synced while your face is on screen, with the original voice talking off camera during the rainbow shot.
- The original audio, uncensored. Contains explicit language.
- People or pets.
- Vertical 9:16, ready for TikTok, Reels and Shorts.
What the Video Actually Is
The clip comes from TikTok creator @eddie_11king, posted on December 5, 2025 and captioned along the lines of "Bro was acting gangster… until he saw a rainbow." It has 725K+ plays, 124K+ likes and 18K+ saves.
The joke is the turn. He opens wide-eyed and dead serious, telling the camera how hard he is, then admits he just saw a rainbow. He spins the phone round to prove it, keeps narrating over a long shot of the sky, and swings back so close that his face fills the whole frame. By then the tough act is gone: he concedes he isn't that tough after all, the rainbow is too beautiful, and his mind has already wandered to what's waiting at the end of it.
Why it suits AI: the clip only works if it's your face doing the sudden softening. A rainbow on demand is the other hard part. One photo covers both.
The 13 Seconds, Shot by Shot
| Time | Camera | What happens |
|---|---|---|
| 0.0s | Selfie close-up, phone held low | You stare into the lens, wide-eyed and serious, acting hard, and start talking |
| 2.1s | Whip pan | The phone swings round and up off your face, a blur of grass |
| 2.4s | Rear camera, drifting | The lawn, the street lamps and the low buildings, a rainbow rising on the right |
| 4.5s | Tilted up at the sky | The phone follows the arc into the pale grey sky, past the power lines |
| 6.0s | Dropping back down | The rainbow stands over the treetops and brick buildings; you're still talking off camera |
| 8.0s | Handheld | A stranger in a white T-shirt wanders across the lawn |
| 8.4s | Whip down and back | The phone sweeps past the grass and swings back to your face |
| 8.8s | Extreme selfie close-up | Your face fills the frame, eyebrows to chin, in front of a red-signed shop: soft, amazed, won over, to the last word |
There are no cuts. Every change of scene is the phone whipping around, which is why it feels like a real clip.
Your Look
You keep your own face, hairstyle and outfit from your photo. Only your face and shoulders show in the selfie shots, so the photo's face matters far more than the outfit. Glasses or a hat stay on only if your photo shows them.
For a pet, upload a photo of the animal on its own. If you're in the photo holding your dog, the template plays the dog. The pet's own face fills the frame, and it keeps its own ears, snout and fur.
Use a sharp, front-facing photo with good light and no sunglasses. The last four seconds are your face filling the whole frame, and that's the shot people screenshot.
The Fastest Way
I Just Seen A Rainbow
The tough-guy rainbow selfie, starring you or your pet. One photo, 13 seconds, every word lip-synced.
One photo, 150 credits (about $7.50), no prompt. Vertical 9:16, 480p, 13 seconds, with sound. Find it in Memes & Skits.
Build It Yourself
1. Drive the motion from the clip. A selfie whip pan is very hard to describe in words. Feed the video model a depth map of the original clip, with the original audio inside it, so it copies the exact camera moves and the mouth timing without copying the creator's face.
2. Make one character sheet. One image with the empty town under the rainbow on top, and you from four sides below. It tells the model what the place and you look like without fixing the framing. (Why a sheet instead of a storyboard: Character Sheet vs Storyboard Grid.)
A photorealistic vertical 9:16 character reference sheet in two bands separated by thin black lines, no text. Top band: a small American town on a grey overcast evening just after the rain, a big bright complete rainbow arcing across the sky over a wide green lawn, a quiet road with black street lamps and power lines, low brick buildings and the beige front of a small shop with a long red sign band, nobody in it. Bottom band: the person in the reference photo, a full-figure turnaround in four equal panels (front, three-quarter, side, back), relaxed, mouth closed, the town softly out of focus behind. They keep their exact face, hairstyle, age, build and clothes. Real photography, natural skin and fabric texture.
3. Write the video prompt with times. Say the sheet is only for looks, list each moment with a timestamp, and give the windows where the mouth talks (0 to 2.1 seconds and 8.8 to 13 seconds) and where no one on screen speaks.
Common Mistakes
Cropping the original to vertical. The source is 4:5. Cropping it cuts off the sides of the face in the close-up; extending the frame top and bottom keeps the whole head.
A mouth that talks during the rainbow shot. The voice keeps going while the camera films the sky. Say that nobody on screen talks between 2.1 and 8.8 seconds, including the stranger.
The rainbow fading or vanishing. Ask for a big, bright, complete rainbow in the character sheet, and it holds through the whole sky shot.
A soft, generic face at the end. The closing close-up is extreme, so use a sharp photo and say that the likeness matters most there.
Related Reading
- How to Make the Nazeer Bakra Goat AI Video: another one-take selfie meme, lip-synced
- How to Make the 'I Just Beat The Case' AI Video: one photo, one line, pure attitude
- How to Make the 'Deep Down In My Heart' AI Video: another tough exterior that cracks on camera
- How to Make the 'It's My Life' Dog AI Video: your pet sings at the wheel
- Character Sheet vs Storyboard Grid: which reference to give Seedance, and when
- Viral AI Video Trends 2026: every trend we've turned into a template this year