Quick answer
Starrd's Island Boy template turns two photos into the viral Island Boy pool meme. Photo 1 sits on the stone edge of a backyard pool with their feet in the water and sings the first half of 'Island Boy' ('Because I'm an island boy...') with a sly, deadpan face. Then the handheld camera swings over to photo 2, who has been standing waist-deep in the pool fully clothed the whole time, throws their head back on 'ayyy' and sings the rest with big rolling hand gestures. One continuous take, every line lip-synced to the song. Everyone keeps their own face, hair and outfit. 14 seconds, vertical 9:16, Seedance 2.5, 480p, 150 credits (about $7.50).
One of you sits on the pool edge and sings "I'm an island boy" with a straight face. The camera swings over, and the other one has been standing waist-deep in the pool, fully clothed, the whole time. Two photos put you and a friend in the viral Island Boy pool meme. Here it is:
Island Boy
Two photos. One sings from the pool edge, one from waist-deep in the water, fully clothed. 14 seconds.
What You Get
- Two parts. Photo 1 on the pool edge sings first; photo 2 in the water takes over.
- One handheld take. The camera swings to the singer in the pool and back, like the original phone video.
- Every line lip-synced. Breaths and handover in the same places as the song.
- Your own look. Your faces, your hair and the clothes from your photos, at a turquoise backyard pool with banana plants behind.
- 14 seconds, vertical 9:16, with the song, ready for TikTok, Reels and Shorts.
What the Video Actually Is
The song is "Island Boy" by Kodiyakredd and Flyysoulja, the Island Boys (2021), whose "I'm an island boy" hook turned into a meme that never quite went away. The pool version is the one that keeps getting remade: a handheld phone video at a backyard pool, one friend sitting on the stone edge with their feet in the water, singing the first half deadpan, while the other stands waist-deep in the pool in full clothes, still and silent, until the camera swings over and they take the second half with a huge grin and rolling hands.
Why it suits AI: the joke needs a pool and a friend willing to walk into it fully dressed. Two photos get you both without anyone getting wet.
The 14 Seconds, Beat by Beat
| Time | Shot | What happens |
|---|---|---|
| 0:00 | Medium-wide two-shot | Photo 1 sits on the edge, leaning toward photo 2; photo 2 stands waist-deep, still, mouth closed |
| 0.7s | Same | Photo 1 sings "Because I'm an island boy", sly and deadpan |
| 2.9s | Same | "...and I've been trying to make it" |
| 4.9s | Same | "Oh, I'm an island boy", then photo 1 goes quiet |
| 7.0s | Camera swings right | Photo 2 throws the head back on "ayyy" with a huge grin |
| 7.9s | On photo 2 | "I'm a just island boy", twice, with rolling, pointing hands |
| 10.5s | Camera eases back | Photo 2 sings the last line, hand out; photo 1 watches with a half-smile |
Your Look, Their Moves
- Your own outfits. A hoodie, a T-shirt and jeans, a suit: whatever is in each photo goes to the pool, and photo 2 wears it into the water.
- Your own hair. Long hair, short hair, beard or not, everyone keeps their own.
- The pool stays the same. Clear turquoise water, a cream stone edge and a wall of banana plants in soft daylight.
Photo 2 has the bigger moment: the grin, the "ayyy" and the hand moves. Put the friend who hams it up second, and the one with the best deadpan first.
The Fastest Way
Island Boy
Two photos: sing it from the pool edge, then from waist-deep in the water. One take, 14 seconds.
Two photos, 150 credits (about $7.50), no prompt. It renders on Seedance 2.5 at 480p, vertical 9:16, 14 seconds, with the song on top.
Build It Yourself
1. Turn the clip into a depth map. Run the clip through a depth-estimation model (Depth Anything V2 works). The depth map keeps the poses, the hand moves and the handheld camera, and drops the faces, the clothes and the colour, so nobody from the original shows up in your video.
2. Make a storyboard. A vertical 2×3 sheet of six stills of your two subjects at the pool, generated from your photos.
A photorealistic VERTICAL 9:16 storyboard sheet of SIX film stills in a grid of two columns and three rows, separated by thin black lines, with no text. All six stills are frames from one handheld phone video at a small backyard swimming pool with clear turquoise water, a pale cream stone edge and a wall of green banana plants behind, in soft daylight. The subject from reference image 1 always SITS on the pool edge on the LEFT with legs and feet in the water; the subject from reference image 2 always STANDS WAIST-DEEP in the water on the RIGHT, fully clothed and soaked from the waist down; each wears exactly the outfit from their own photo. Top row and middle-left: medium-wide two-shots, the left subject singing with a sly deadpan face, the right one still and listening. Middle-right and bottom-left: closer on the right subject, grinning and singing with rolling, pointing hands. Bottom-right: the two-shot again, the left subject watching with a half-smile.
3. Drive it with the depth map, with the song inside. Feed the storyboard, your photos and the depth video (with the song as its audio) to a model that copies motion from a reference video (we recommend Seedance 2.5). Then write what it can't read from a depth map:
- Who is who: photo 1 always seated on the left, photo 2 always standing in the water on the right.
- The handover: when photo 1 stops singing and photo 2 starts, to the tenth of a second.
- The faces: the deadpan, the grin and the half-smile. The faces in the original are small, so the expressions won't come through the depth map on their own.
Lay the song back on top at the end: Seedance can drop or change music on its own, so mux the track yourself.
Common Mistakes
Driving with the raw clip. The original faces come straight through. Use the depth map.
Letting photo 2 climb out. Models like to put both people on dry land. Say photo 2 stays waist-deep in the water the whole time, and photo 1 never stands up.
Both singing at once. Without explicit windows, both mouths move from the first second. Give each singer their window and say the other keeps their mouth closed.
Dropping the swing. The camera pans right onto the singer in the pool around 7 seconds and back around 10.5. If the prompt says "static", the reveal is lost.
A silent driver. With no song inside the depth video, the lip sync drifts. Put the track in the driver as well as on top.
Related Reading
- How to Make the Careless Whisper Car AI Video: another deadpan friends meme, in one take
- How to Make the 'Beans Greens Potatoes Tomatoes' AI Video: a lip-synced meme line, word for word
- How to Make the 'I Need A Girl' AI Dance Video: a lip-synced intro, then a two-friend dance
- How to Make the 'Closer' by Ne-Yo AI Dance Battle Video: two friends, one song, one car park
- Viral AI Video Trends 2026: every trend we've turned into a template this year
