Homelander Milk AI Prompt (Copy-Ready): The Whole 15-Second Spill
This Homelander milk AI prompt gives you the copy-ready prompt for the 15-second kitchen clip: both people, every beat with seconds, the song, and one photo.
The Seadanse Team13 min readThe team behind Seadanse. We run the models we write about, and we publish the specs, prices and limits the marketing leaves out.
TL;DR — The Homelander milk AI prompt below is a copy-ready prompt for the whole 15-second kitchen clip: who does what in all 7 beats, the kitchen, the hero's costume and the song. It also has a line for one photo and tells the tool what never carries over from your reference clip. To skip the prompt, Seadanse's Homelander Milk AI page makes the same clip from 1 or 2 photos, with the song in the file.
Under NS7's Homelander milk clip on TikTok, a viewer asked "con que IA lo hacen" ("what AI do they make it with"). And 109 people liked it. Under NS7's October 20 Homelander milk clip, another viewer asked "¿Cómo se hace este efecto?" ("how do you do this effect?"), and 405 people liked it.
Nobody had posted a prompt or a step-by-step for this clip when we looked on October 29, 2026. A one-line remix instruction leaves the tool guessing who spills and who stares.
Here's a full prompt you can paste, how it's built, how to use it, and five habits that keep it on the beat.
What Is the Homelander Milk AI Prompt?
The Homelander milk AI prompt is the written instruction a video tool reads to remake the kitchen clip with your photos. Here it is, for two photos:
People:
Person A is the one in photo 1 (a person, a cartoon or game character, or a pet): [describe A's face, hair and outfit]. Person A is the one who pours the hot water, fetches the milk, pours it all over the counter and shrugs.
Person B is the one in photo 2 (a person, a cartoon or game character, or a pet): [describe B's face and hair only]. Person B is the caped hero who makes the cocoa, lifts the mug and stares.
Person A keeps their own face, hair, body and outfit from photo 1 in every frame. Person B keeps their own face, hair and body from photo 2 in every frame, and always wears the same costume: a plain charcoal-grey bodysuit with a round neckline, a long red cape and red leather gloves, with no emblem, logo or gold trim. Person A's face never appears on Person B, and Person B's face never appears on Person A. Nobody else is in the kitchen.
If photo 1 shows a cartoon, anime or game character, the whole video is in that style. If it is a real photo, the whole video is photoreal. A pet stays a real animal of its own breed, face, ears and coat, standing at the counter and acting like a person.
Set:
A home kitchen in warm daylight: honey-oak wooden cabinets under a floral wallpaper border, a white ceiling with a round recessed light, a steel sink on the left, and a speckled beige granite counter in the foreground with a red mug, a glass jar of cocoa powder and a small pile of spilled cocoa on it. Person A stands behind on the left; Person B stands in front on the right. One continuous shot from a low camera at counter height with a slight drift, vertical 9:16, no cuts.
Reference clip:
Your reference clip is the 15-second kitchen clip. It is for motion and timing only: carry over how the scene is filmed and performed, every movement, prop, splash, mouth movement, facial expression, gaze and the timing. Take no face, hair, body or clothing from it: the two people in it do not appear anywhere in the video, not even for a moment, and neither person keeps the original face or hair.
Beats:
0-2.5 s: Person A pours hot water from a steel kettle into the red mug while Person B spoons cocoa from the glass jar into it.
2.5-7 s: Person A puts the kettle back and stands back; Person B scoops more cocoa and looks into the mug.
7-9.5 s: Person A turns away to the back counter; Person B turns their head toward Person A as if saying something.
9.5-10.2 s: Person A comes back holding a white plastic milk jug; Person B smiles down at the mug and lifts it off the counter.
10.2-11.3 s: Person A pours the milk straight onto the bare counter where the mug stood, and it splashes into a wide white puddle that spreads toward the jar.
11.3-12 s: Person B's smile drops; they look down at the puddle, then slowly turn toward Person A.
12-15 s: Person B holds the red mug in both gloved hands and stares at Person A in profile, perfectly still, until the last frame; Person A shrugs, opens their hands and makes an excuse.
Sound:
The reference clip's song plays under the whole video from the first second. No captions, logos or text on screen.The parts in square brackets are yours to fill in, one short description per person. Everything else stays as it is.
How the Homelander Milk AI Prompt Is Built
The Homelander milk AI prompt is built in five labelled parts, and the longest is the beat list.
| Part | What it tells the tool | Why it's there |
|---|---|---|
| People | Who's in photo 1 and photo 2, who spills and who stares | So the right face does the right job |
| Set | The kitchen, the counter, the red mug, the camera | So the scene matches the clip |
| Reference clip | Copy the moves and timing, nothing else | So nobody from the clip shows up |
| Beats | Seven beats with their seconds | So the spill lands on the same beat |
| Sound | The clip's song from the first second | So the stare plays over the song |
How Do You Use the Homelander Milk AI Prompt?
To use the Homelander milk AI prompt, you need a video tool that takes two photos and one reference clip together, a copy of the 15-second kitchen clip, and your photos. A tool with no reference clip has nothing to copy the moves and timing from, so the prompt's written for a tool that takes one.
Here's what you need:
- The tool: a video tool that takes two photos and one reference clip together, in one go.
- Photo 1 and photo 2: photo 1 is Person A, the one who spills; photo 2 is Person B, the caped hero.
- The reference clip: a copy of the 15-second kitchen clip, the one the tool copies the moves and timing from.
- The fill-ins: one short description per person in the square brackets, for example hair colour and what they wear.
- The song: if your tool makes silent video, add the song in an editor afterwards.
If you'd rather skip the setup, the Homelander Milk AI generator does it from your photos.
Use the Homelander Milk AI Prompt With One Photo
The Homelander milk AI prompt works with one photo if you swap the Person B line for the partner line below.
Person B is a made-up adult man aged 25 to 32, never a real or famous person and not resembling any actor: a friendly, likeable face, an athletic build, short dark-brown or black hair in a natural textured cut, brown, hazel or green eyes, and light stubble or a short trimmed beard. His hair colour or style is clearly different from Person A's. He wears the costume described above and no text or logos. Keep the same Person B, with the same face and hair, in every frame.The made-up partner is an adult man aged 25 to 32 with short dark hair, light stubble or a trimmed beard, wearing the hero's costume. Photo 1 can be you, a cartoon character or your dog, and "A pet stays a real animal of its own breed, face, ears and coat, standing at the counter and acting like a person."
Habits That Keep the Homelander Milk AI Prompt on the Beat
Five habits keep the Homelander milk AI prompt on the beat, and you can see each one in the prompt above.
Name Who Spills and Who Stares
Person A and Person B need separate jobs from the start, so you'll name them in the first lines. In our prompt, Person A is the one who "pours the hot water, fetches the milk, pours it all over the counter and shrugs". Person B is "the caped hero who makes the cocoa, lifts the mug and stares".
If you leave these jobs unassigned, the tool mixes up the actions. That's why the prompt says "Person A's face never appears on Person B" and makes sure "Nobody else is in the kitchen".
Put the Costume on the Hero Only
Person B "always wears the same costume": "a plain charcoal-grey bodysuit with a round neckline, a long red cape and red leather gloves". The costume belongs on Person B and nobody else.
Person A keeps their own outfit from your photo. That costume's written out in the prompt, not taken from the clip, so Person A doesn't end up in a cape.
Time All Seven Beats to the Clip
Every beat needs its exact seconds so the spill and the stare land right on time. We've written down the clip's own timing across seven beats:
- Beat 1 (0-2.5 s): Person A pours hot water from a steel kettle into the red mug; Person B spoons cocoa into it.
- Beat 2 (2.5-7 s): Person A puts the kettle back and stands back; Person B scoops more cocoa.
- Beat 3 (7-9.5 s): Person A turns away to the back counter; Person B turns to say something.
- Beat 4 (9.5-10.2 s): Person A comes back with a white milk jug; Person B smiles and lifts the mug.
- Beat 5 (10.2-11.3 s): Person A pours the milk straight onto the bare counter, and it splashes into a wide white puddle.
- Beat 6 (11.3-12 s): Person B's smile drops; they look down at the puddle, then turn to Person A.
- Beat 7 (12-15 s): Person B holds the red mug and stares, perfectly still; Person A shrugs and makes an excuse.
Say What Never Carries Over
When a video tool reads a reference clip, it'll copy what it sees unless you tell it where to stop. That's why our prompt makes the boundary clear.
The prompt states: "It is for motion and timing only". It also adds: "Take no face, hair, body or clothing from it", "not even for a moment".
Keep the Song Under the Whole Clip
The sound line states: "The reference clip's song plays under the whole video from the first second." The song under the clip is Julio Iglesias's "Con la misma piedra", and you can listen to the official recording of Julio Iglesias's Con la misma piedra.
If your tool makes silent video, you'll add the track in an editor afterwards. The prompt also notes: "No captions, logos or text on screen."
Our Homelander Milk AI Prompt vs a One-Line Remix
A TikTok remix of the clip uses a one-line instruction, and its caption starts "Replace the person in the drama with …". That instruction names the people only by where they stand ("being the person on the left").
That short line doesn't have beats, seconds, a costume line or a song line. It leaves the video tool to guess when the milk spills and who wears the cape.
| One-line remix instruction | Our Homelander milk AI prompt | |
|---|---|---|
| Who's who | "the person on the left" | Person A spills, Person B stares |
| The costume | Not mentioned | On the hero only, written out |
| The kitchen | Not mentioned | Written out |
| Timing | Not mentioned | 7 beats with seconds |
| The song | Not mentioned | Under the whole clip |
| One photo | Not mentioned | A made-up partner line |
Our prompt adds who's who, the costume on the hero only, the kitchen written out, all seven beats with their seconds, the song and a one-photo line.
Can You Make a Homelander Milk AI Video Without a Prompt?
Yes, you can make a Homelander milk AI video without a prompt on Seadanse's Homelander Milk AI, where the beats, the kitchen, the costume and the song are built in.
Here's how you make it:
- Add photo 1: upload the person, character or pet who spills the milk.
- Add photo 2 or leave it: upload the person for the caped hero, or leave it empty for a made-up partner.
- Pick 480p, 720p or 1080p: choose the export quality you want for your MP4.
- Press Create: make the video without typing any instructions.
On Seadanse pricing, one 720p video costs $9.90, paid once without a recurring membership, and you don't need to sign up.
You can also read how to make a Homelander milk AI video step by step to see the full four steps.
Homelander Milk AI Prompt FAQ
Is this the prompt Seadanse's page uses?
It's written from the same beat list our page is built on: who spills, who stares, the kitchen, the costume and the song. On our page you don't see or write a prompt. You add photos and press Create.
Which AI video tools accept the Homelander milk AI prompt?
The prompt works with any video tool that takes two photos and one reference clip together in one go. You can see the Homelander AI video generators compared to check how different tools handle the inputs.
Does the Homelander milk AI prompt keep the song?
Yes, the sound line asks for the clip's song under the whole video. The song's Julio Iglesias's "Con la misma piedra", and on Seadanse it plays from the first second. Read where the Homelander milk AI meme comes from for the story behind the song.
Can I use the Homelander milk AI prompt for my dog or a cartoon character?
Yes, you can. Pets and cartoon characters both work with this prompt. A cartoon or anime character sets the visual style for the entire clip, while a pet stays a real animal standing at the counter.
Why doesn't the prompt name Homelander?
It describes what you see, not who it is, so it doesn't need a name. The prompt describes a caped hero in a grey suit and a red cape, a kitchen and a red mug, so your two photos set the faces.
Ready to Skip the Homelander Milk AI Prompt?
You can make a Homelander milk AI video right now on Seadanse from one or two photos, with every beat and the original song included. One video costs $9.90 at 720p, with no watermark and no prompt to write.
Written by The Seadanse Team, the team that built the Homelander Milk AI page.
Related Articles
- How to Make a Homelander Milk AI Video From 1 or 2 Photos (4 Steps) — the photos, the quality, Create.
- Homelander Milk AI Meme: The Spilled-Milk Kitchen Clip, Explained — what the clip is, where its beats come from, and why the stare works.
- Homelander Milk CapCut Template: Which One Makes the Kitchen Clip — the kitchen template, the 2024 look-alikes, and five fixes.
- Best Homelander AI Video Generators (2026): 8 Ways to Make the Milk Clip — the sites and apps side by side on photos, length, sound and price.
