Kling 4.0 + GPT Image 2.6 Flare - funny ads from four famous print ads, starring your own clips
Turn four famous print ad gags into funny video ads using your own 3 second phone clips
Disclaimer: This is an independent creative test made with AI, for demonstration only. It isn't an ad for any brand shown or named, and isn't affiliated with, sponsored or endorsed by them. All trademarks belong to their owners
This is the video I posted on Instagram, here's the link if you want to watch it with sound.
In this post you'll learn how to:
- turn a famous print ad gag into a 3 second video starring you
- make your head so heavy you have to drag it up the stairs by your neck
- blow a bubblegum bubble that kisses you back
- walk on the tea inside a giant teacup and cry a flood that fills the frame
- leave a clean spot in every shot so you can paste the real product in later
I'm taking famous print ads and turning them into funny video ads where I'm the guy in the ad. A painkiller ad, a bubblegum ad, a tea ad and a paper towel ad. None of them are clients, this was just a test. Each one starts as a plain 3 second clip of me doing something boring, like walking up stairs or sitting on a bench.
Here's my exact process.
1. Pick a print ad with one physical joke
First, I look for print ads where the whole joke is one thing you can see. A headache that makes your head heavy. A bubble that kisses you. A teabag so big you carry it like a shopping bag. Crying so hard you flood the place.
Then I ask myself what I'd be doing in the real world when that joke happens. Heavy head means walking up stairs. Kiss means sitting on a bench looking at the empty seat. That's what I film.
2. Film a boring 3 second clip
I film each one on my phone, in landscape, 3 to 5 seconds, one simple action. No acting needed for most of them, the AI does the funny part.
By the way, keep your clip at 3 seconds or more. My bench clip came out at 2.96 seconds, which is too short for the video edit. So I slowed it to 90% speed with smooth slow motion, and you basically can't tell.
3. Make the still from the very first frame
Now I grab the first frame of my clip and turn that into the edited still. Your video starts from that frame, so if you build the still from the middle of the clip, the video starts somewhere you weren't, and it gets weird.
I keep the prompt simple: keep the photo exactly as it is, same camera, same framing, same me, and change one thing.
4. The headache: drag your own giant head
For the painkiller ad, my head gets so big and heavy that it falls off my shoulders and I drag it up the stairs by my neck.

Exact prompt
Keep the first reference photo exactly as it is, pixel for pixel: same camera, framing, crop, stairway, sign, buildings, palms, golden-hour light, the man's body and pose, the long neck, and the giant upside-down head in the foreground with its face, glasses, hair and expression. Change only ONE thing: remove the small normal-size head sitting on the man's shoulders. Above his t-shirt collar there is no head at all, exactly like the empty shoulders in the second reference: just the collar, the base of his neck, and the stairs and sky behind. His only head is the giant one in the foreground. THE VIBE: A real iPhone photo of a surreal headache ad: a man climbs a stone stairway at golden hour while his own enormous head, upside down, is dragged along the steps behind him by his long neck. Photographic. THE SUBJECT: The man exactly as in the first reference from the shoulders down: black t-shirt, dark jeans, sneakers, same step and arm position. Nothing on his shoulders. His neck rises from the back of his collar and curves back down the steps to the base of the giant head, exactly as in the first reference, as one continuous neck. THE SETTING: Identical to the first reference. Where the small head was, show the stone stairway and the pale-blue sky that would be behind it, matching the surroundings seamlessly. THE KEY DETAILS: Exactly one head in the image: the giant upside-down head in the foreground. No second head, no hair or ears on his shoulders. Nothing else changes. THE LIGHT: Identical to the first reference: low warm golden-hour light from behind-left, pale-blue sky, warm stone. Natural iPhone colour. THE MOOD: Absurd and relatable: a headache makes your head this heavy. THE COMPOSITION: Identical to the first reference, no zoom, no reframing. THE CAMERA: iPhone main camera, landscape, handheld eye-level, sharp, natural.
The trick that made it look natural was flipping the head upside down, so my neck joins it right under the jaw, where a real neck goes. Then I made sure there was nothing left on my shoulders, otherwise you get two heads.
Exact prompt
Create a new video that is @video_1 frame for frame: the same camera, the same framing, the same stairway and golden-hour light, and the man's body with exactly the same motion from the very first frame to the last. The only change: he has no normal head on his shoulders. His own head is gigantic and far too heavy, lying upside down on the steps behind him, and his long neck drags it up the stairs behind him as he climbs, exactly as in @image_1. A surreal headache ad, real iPhone footage, deadpan. [SCENE 1] The video starts exactly on the first frame of @video_1, with the man exactly where he is there, and his position in frame, every step up the stairs, his arm swings, his timing, clothes and sneakers all come only from @video_1, unchanged. Take NOTHING about his position, pose or placement from @image_1. @image_1 is only the look of the change: nothing above his t-shirt collar for the whole clip, no head, no hair, no ears on his shoulders in any frame; his neck rises from the back of his collar and runs back down the steps toward the camera to the base of one gigantic upside-down copy of his own head, which lies on its back on the steps behind him, chin toward him, crown toward the camera, face up with glasses, eyes squeezed shut in pain, exactly as in @image_1. There is only ever one head in the frame: the giant one. As he climbs, the neck pulls taut with each step and the heavy head drags along behind him: it scrapes along the stone, catches on each step edge, then lurches and bumps up onto the next step with a heavy thud, wobbling and squashing slightly on impact, always trailing a little behind his body. The face stays wincing, the glasses jostle askew on each bump. The neck stretches and slackens with his rhythm and stays one continuous neck from collar to head. Same low warm golden-hour light from behind-left as @video_1, the head and neck rim-lit gold, a moving contact shadow under the head; the sign, palms and buildings exactly as in @video_1.
5. The bubblegum: a bubble that kisses you back
For the bubblegum ad, I'm sitting on a bench and blow a bubble shaped like a girl's head that kisses me. I asked for a ponytail made of the same pink gum, which is what made it read as a girl.

Exact prompt
Keep the first reference photo exactly as it is: same camera, framing, crop, plaza, bench, daylight, the man exactly as he sits with his face, glasses and clothes, no aging, and the pink bubblegum bubble face kissing him at the same size and position. The second reference is the original photo, for the man's exact face and the plaza. Change only one thing: the bubblegum head is a girl with a ponytail. THE VIBE: A real iPhone photo of a bubblegum print ad come to life on a park bench: a man blows a bubble of pink gum that has formed into a girl's head kissing him, complete with a ponytail made of the same pink gum. Playful, deadpan, photographic. THE SUBJECT: The man exactly as in the first reference: black t-shirt, navy trousers, glasses, on the bench, arm along the backrest, lips puckered and joined to the bubble's lips. Preserve his face exactly, no aging. THE SETTING: Identical to the first reference. THE KEY DETAILS: The bubblegum head as in the first reference, now clearly a young woman's head in profile, all made of glossy semi-translucent pink bubblegum: delicate feminine features, closed eye with a soft lash line, small nose, full puckered lips pressing into his in a kiss. Sculpted gum hair pulled back smoothly from her forehead into a high ponytail at the back of the head: the ponytail is a thick, bouncy swoop of the same pink gum, gathered by a slightly thicker gum band, curving out and hanging down off the back of the head on the left side, with soft strand-like ridges and a stretchy, glossy gum texture. Everything is one continuous pink gum material, same translucency and sheen, plaza faintly visible through the thinner parts. No real hair, no text, no product, no logo. THE LIGHT: Identical to the first reference: soft late-afternoon daylight, a bright soft highlight on the top of the gum head and along the ponytail's curve, highlights on the brow, nose and lips, a gentle pink glow on his cheek, soft shadow underneath. Natural iPhone colour. THE MOOD: Funny, sweet, made-for-kissing. THE COMPOSITION: Identical camera and crop to the first reference, no zoom. The man stays where he sits; the gum girl's head to the left of his face, the ponytail fully in frame. THE CAMERA: iPhone main camera, landscape, handheld eye-level, sharp, natural.
In the video I told it to keep my face almost still and let the gum do all the moving: the bubble sways in the breeze and the ponytail swings a beat behind.
Exact prompt
Create a new video that is @video_1 frame for frame: the same camera, framing, plaza, bench, daylight, and the man with exactly the same body motion and timing from the first frame to the last. The only change: from his first frame on, his lips are puckered and joined to a big pink bubblegum bubble shaped like a girl's head with a ponytail, kissing him, exactly as in @image_1. Real iPhone footage, playful, deadpan. [SCENE 1] The video starts on the first frame of @video_1 with the man exactly where he sits, arm along the bench back; his body, arm, posture, clothes and glasses all follow @video_1. His face stays almost still the whole clip: head held steady in profile toward the bubble, lips puckered and locked to the bubble's lips, eyes calm, only a slow natural blink and slight breathing, no turning away, preserve his face exactly, no aging. @image_1 is the look of the bubblegum girl only: a large glossy semi-translucent pink gum head in profile, closed eye, small nose, puckered lips pressing into his, a high ponytail of the same pink gum with a gum band. The bubble stays attached to his lips for the entire clip and never pops, detaches or disappears. The bubble is alive with soft motion: it gently sways and bobs in the light breeze, drifting slightly up and down and side to side around the fixed point at his lips, its thin membrane softly wobbling and flexing, the sheen and highlights sliding across its surface as it moves. The ponytail swings and bounces with a slight delay after each sway, its tip lagging and settling, swaying side to side. Its lips stay pressed to his, puckering slightly into the kiss. Same soft late-afternoon daylight as @video_1, highlights moving on the gum, a gentle pink glow on his cheek; the person walking in the background and the plants move as in @video_1.
6. The tea: walk on a giant teacup
For the tea ad, I'm balancing on a ledge in a park. The ledge becomes a giant white teacup full of tea, I walk on the tea, and I carry a giant teabag.

Exact prompt
Keep the reference photo exactly as it is: same camera angle, framing, crop, park, palm trees, hedges, white buildings, sunlight, and the man's exact position at the right of the frame, his stance, clothes and sneakers. His arm on the left side of the image stays held out for balance exactly as in the reference, hand open and empty. Change only two things: the stone ledge becomes a giant teacup full of tea that he stands on, and his hand on the right side of the image grips a colossal teabag. THE VIBE: A real iPhone photo of an ambient ad come to life in a sunny palm park. A man balances on the surface of the tea inside an enormous white porcelain teacup as if it were solid ground, gripping a gigantic teabag. Deadpan, effortless, believable, not a render. THE SUBJECT: The man from the reference, preserved exactly where he stands at the right of the frame: navy t-shirt, dark navy jeans, black and white sneakers, same stance, framed from the chest down. His arm on the left of the image out for balance, empty hand, exactly as in the reference. His arm on the right of the image hangs down, hand closed in a tight fist directly around the folded top of the giant teabag, the bag's top right under his knuckles, no strings, no gap. THE SETTING: The stone ledge is replaced by an ENORMOUS white glossy porcelain teacup: its rounded rim runs edge to edge across the whole frame, its curved white side fills the bottom of the frame and continues off the bottom edge, and a thick curved porcelain teacup handle sticks out from the cup's side on the right and runs off the right edge of the frame, so it clearly reads as a giant teacup. Filled to the brim with clear warm amber-brown tea. Same park behind as the reference. THE KEY DETAILS: His sneakers stand ON the tea surface, not sinking: concentric ripples spreading from each sole, a glossy dimple at the contact point, reflections of his legs, palms and sky in the tea. The teabag: a GIGANTIC white paper tea bag, the classic flat rectangle of white filter paper with its top folded into a point, about as tall as from his hip to just above the tea and as wide as his body, the biggest object after the cup. It hangs slightly in front of his legs toward the middle of the frame so it is fully in frame and not cut off by the edge. Its lower half shows dried green tea leaves packed behind translucent white filter-paper mesh. From the folded top, on a short white string, hangs a separate square paper tag in a yellow-to-green gradient with a red-and-yellow sunburst oval logo reading "Lipton" in white, and green lowercase text "clear" above "Green". Its bottom edge hovers just above the tea, not touching it, with its shadow and reflection on the surface. THE LIGHT: KEY: the same bright late-afternoon sun as the reference, from the upper left, hard; warm highlights on the porcelain and the bag, soft blue-grey shadow on the far sides, real cast shadows of him and the bag on the tea. FIELD: the same bright, saturated park as the reference. Natural iPhone colour. THE MOOD: Clever, clean, satisfying ambient-advertising gag. THE COMPOSITION: Exactly the reference's camera, distance and crop, no zoom. The man stays at the right of the frame. The cup dominates the lower frame edge to edge with its handle exiting right; the bag hangs in front of his legs, fully visible. THE CAMERA: iPhone main camera, landscape, handheld eye-level, sharp, natural.
One tip: if you're holding something, say in the video prompt that it stays in your fist from the first frame to the last. Otherwise the model will happily let go of it halfway through.
Exact prompt
Create a new video that is @video_1 frame for frame: the same man, the same camera, the same framing, and exactly the same motion from the very first frame to the last. The only changes: the stone ledge he walks on is now an enormous white porcelain teacup full of amber tea, so he walks on the tea's surface as if it were solid ground, and he holds a giant teabag in one hand for the ENTIRE clip, exactly as in @image_1. Bright sunny palm park, real iPhone footage, deadpan. [SCENE 1] The video starts exactly on the first frame of @video_1, which matches @image_1: the man at the right of the frame, one arm out for balance, the other hand gripping the folded top of the teabag. His position, every step, his balance, the turn, the walk back, his timing, clothes and sneakers all come from @video_1, unchanged. @image_1 shows the look of the new objects: the giant white glossy porcelain teacup (rim edge to edge, thick handle running off the right edge, full of clear amber-brown tea) replacing the ledge for the whole clip at a fixed size and position, and the giant white filter-paper teabag with dried green tea leaves in its lower half and the yellow-to-green "Lipton clear Green" tag on a short string. THE BAG IS HELD THE WHOLE TIME: from the first frame to the last, his fingers stay closed tightly around the bag's folded top, in the same hand as in @image_1; he never opens that hand, never lets go, never switches hands. Wherever that real arm and hand move in @video_1, the bag's top moves with them, through the whole turn. It never floats, never drifts from his hand, never disappears. It hangs from his fist with real weight: swings with his arm, dips on each footfall and springs back, lags a beat behind his body when he turns, its bottom skimming just above the tea, the tag swinging. His sneakers land ON the tea and never sink: each step sends concentric ripples across the tea that keep travelling, his legs reflected wobbling in it. Same bright late-afternoon sun from the upper left as @video_1, real moving shadows of him and the bag on the tea, background people and palms moving as in @video_1.
7. The tears: cry a flood
For the paper towel ad, I'm just sitting there. The AI turns my face into a full ugly cry, with two giant streams of tears pouring out of my eyes into water that fills the bottom of the frame.

Exact prompt
Keep the first reference photo exactly as it is: same low camera angle, framing, crop, hedge, trees, blue sky, sunlight, the concrete ledge, and the man exactly where and how he sits, his head position and turn, his arm along the ledge, his navy t-shirt and glasses. The first reference sets his head size, head shape and proportions, never larger. The second reference is only for his facial features and glasses, never for the size or scale of the head, and nothing else from it. Change two things: he is ugly-crying, and an absurd flood of tears pours from his eyes and fills the bottom of the frame with water. THE VIBE: A real iPhone photo of a comedic crying print ad come to life: a man sobbing so hard that huge gushing streams of tears pour out of his eyes and the water has already flooded the bottom of the frame up to his chest. Absurd, funny, photographic, not a render. IDENTITY LOCK: His face preserved exactly as in the first and second references, no aging, no added wrinkles, same beard, same glasses. THE SUBJECT: The man from the first reference in the same pose and head angle, now with a full ugly-cry expression: brow pinched up in the middle, eyes squeezed half shut and glossy, mouth open and pulled down at the corners in a wail, bottom lip trembling and pushed out, cheeks wet. His shoulders slightly hunched. THE FACE: Real skin texture, wet glistening cheeks, reddened eyelids, the expression clearly sobbing, his own face and glasses otherwise exactly as in the references. THE SETTING: Same hedge, trees, sky and ledge as the first reference. Clear water fills the bottom of the frame up to the middle of his chest, covering the ledge and his lower body, the water level flat across the frame edge to edge. THE KEY DETAILS: From under his glasses, two thick, absurdly large streams of tears gush out of both eyes like open garden hoses, arcing outward and down in clear glassy ribbons and splashing into the water below, kicking up spray, droplets and ripples. The tear streams start just below the eyes and the lenses, leaving his eyes and face fully visible, never covering his face. The water surface ripples outward from the splash points, reflecting the hedge, sky and his t-shirt; his t-shirt is dark and wet where it meets the water. No text, no product, no logo. THE LIGHT: KEY: the same bright midday sun as the first reference, from the upper right, hard; sparkling specular highlights on the tear streams, droplets and water, wet glints on his cheeks. FIELD: the same bright blue sky and sunlit hedge as the first reference. Natural iPhone colour. THE MOOD: Over-the-top sad, funny. THE COMPOSITION: Exactly the first reference's camera and crop, no zoom. The man where he sits, the tear streams arcing to either side of his face, the water filling the bottom of the frame. THE CAMERA: iPhone main camera, landscape, low angle, sharp, natural.
I put the water and the streams in the still already, so the video starts mid-flood. Then the prompt just says keep crying, keep pouring, let the water rise.
Exact prompt
Create a new video from @video_1: the same man, the same low camera angle, the same hedge, trees and blue sky, and his real head and body movement and timing from the first frame to the last. The change, exactly as in @image_1: he is ugly-crying the whole clip, two absurd gushing streams of tears pour out of his eyes, and the water they make fills the bottom of the frame up to his chest. Real iPhone footage, over-the-top sad, funny. [SCENE 1] From the very first frame he is already sobbing, the tear streams already pouring and the water already up to his chest, as in @image_1, and it never stops. His head and shoulders move with the real timing of @video_1, plus sobbing: shoulders heaving with each sob, chin trembling, bottom lip quivering and pushed out, mouth open in a wail that opens wider with each sob, brow pinched up, eyes squeezed and glossy, cheeks wet and glistening. His face, beard, hair and glasses stay exactly as in @image_1, no aging. Two thick clear streams of tears gush continuously out from under his glasses, one from each eye, like open garden hoses, arcing out and down in glassy ribbons, pulsing a little stronger with each sob, never stopping, and following his head as it moves. They crash into the water below with constant splashing, spray and flying droplets. The water level slowly keeps rising across the clip, rippling outward from the splash points, the surface sloshing and reflecting the hedge and sky, his wet t-shirt darkened at the waterline. Same bright midday sun from the upper right, sparkling highlights on the streams, spray and water, wet glints on his cheeks; the hedge and trees sway as in @video_1.
8. Paste the real product in
Last step, I bring each clip into the Instagram Edits app, stack it over the original clip so you see the before and after, and paste a photo of the real product into the empty spot. That's why I left the sky, the bench and the frame edges clean in every still.
Why it works
Print ads already did the hard part. Somebody spent weeks finding one image that makes you get the product in a second. I'm just borrowing the joke and putting myself in it, and because the video edit keeps my real movement, it still looks like me filmed it.
Your turn: pick one print ad you remember, figure out what you'd be doing when the joke happens, and film 3 seconds of it.
And if you'd rather not write these prompts yourself, Emero writes them for you and remembers your face and your look between videos, at cost.
Here's the full cut one more time:
Exact prompt
Combine the clips in order
Get one of these every day
I test new workflows daily and send them with the exact prompts, the models and the costs
6,526 generations tested so far
More workflows
See allKling 4.0 edit mode - one phone clip relit in four kinds of cinematic lighting
Relights one phone clip four ways: golden haze, neon ember, spotlight noir and rain city
GPT Image 2.6 Flare, Kling 4.0 · 8 min render
Hook ideas for reels: sweat that turns to steam, a giant selling chains, a crumbling sandcastle and a fire saxophone
Four AI visual hooks from plain phone clips, posted as a before-and-after split
GPT Image 2.6 Flare, Kling 4.0 · 8 min render
AI fat filter, live: a giant belly, a neck stretch with diamond grills, Atlas spinning a rock and a lion ride
A giant belly, diamond grills, a lion ride and Atlas, made from four phone clips
GPT Image 2.6 Flare, Kling 4.0 · 9 min render