I'm So In Love With You AI
RapTwins’s I'm So In Love With You AI is a one-photo meme video generator: pick the two-pose stairwell lip-sync or the four-pose white-floor remix, and get an 8-second MP4 with that edition’s sound of mthu’s “it’s like i’m not even here” already in sync, from 15 credits (about $0.56), 3 free credits at sign-up. No watermark, plus a silent copy.
Updated
Make your I'm So In Love With You AI video
1. Which version
2. Your photo
Which photos work best
JPG, PNG or WebP, 300–6000 px, up to 10 MB. One consenting adult or an original character, facing the camera, with both arms and the hips visible. For the white-floor version a full-body photo works best, because the poses show the legs and feet.
3. Model
8 s · 480p · 9:16 vertical MP4 with the stairwell take's sound, plus a silent copy
Add a photo of one performer first.
Any credits pay, sign-up and check-in credits included. New accounts get 3 free credits, fewer than one video costs (from 15 credits).
A finished stairwell clip: 8 s with its sound · 9:16 · your video keeps these two poses, this stairwell and this vocal, with your face and outfit.
I'm So In Love With You AI video examples
One stairwell take and three white-floor remixes, each with its own sound. The stairwell clip is the two-pose edition your video copies; the three square clips show the four-pose edition with different characters on the same beats. The lyric text in the square clips is part of those examples, not something the generator adds. Tap a clip for sound.
Two poses in the stairwell, with the vocal
The two-pose lip-sync the stairwell edition copies: a woman in a teal sweater mouths the hook with her fists out and elbows bent, then plants both fists on her hips and lifts her chin at about 4.5 seconds. 8 seconds, 9:16, with the take’s own sound.
8 s · with sound · vertical
Four poses on the white floor: armoured explorer
The four-pose white-floor remix with a game character in heavy armour: arms spread wide, a low stance with the forearms in, hands to the hips, then the angled last pose, each about two seconds. 8 seconds, 1:1, with the remix sound and lyric text.
8 s · with sound · square
Four poses on the white floor: cartoon character
The same white-floor remix with a 2D cartoon in a black T-shirt and blue trousers, so you can see the four holds and the hip shifts in a flat drawing style. 8 seconds, 1:1, with the remix sound and lyric text.
8 s · with sound · square
Four poses on the white floor: dark animatronic
The same white-floor remix with a dark animatronic figure: the same four holds on the same beats as the other two square clips. 8 seconds, 1:1, with the remix sound and lyric text.
8 s · with sound · square
What song is in the I'm So In Love With You AI meme?
It is “it’s like i’m not even here” by mthu, a single released on April 10, 2026; the line everyone repeats opens the song, which is why the meme is named after it. It is not Al Green’s “I’m So in Love with You”, The Kid LAROI’s similarly titled track or the other older songs with the same words, which is what a search for the hook alone usually turns up.
Search for mthu and the full track title together to find the sound in TikTok or Instagram. People also call the meme “I’m So In Love With U”, the “fists then hips dance” or, for the cartoon version, “Jiren dancing”.
What does “I’m so in love with you” mean in the meme? It is played as an over-the-top declaration: the performer stares at the lens dead serious, fists out, then plants the fists on the hips as if to say the feeling is obvious. The joke is the dramatic delivery, not a real love message, so it works on friends, pets drawn as characters and game heroes alike.
Does the AI video come with the song, and does the mouth sync?
Yes. The model makes a silent 8-second clip, then the finished MP4 gets the sound of the exact take your poses were copied from, the same 8 seconds, so the fists, the drop to the hips and the hook land together without editing. You don’t pick or upload music, and your own voice is never recorded or cloned.
The two editions carry different audio. The stairwell has that take’s own vocal recording, and the prompt asks the model to mouth the vocal rhythm to the lens, so it reads as a lip-sync. The white-floor remix carries the edit used by the character clips; the mouth matters less there because the joke is the four poses.
Every video also comes with a silent copy, for posting with the app’s own sound instead (see below).
Two poses or four poses: which I’m So In Love With You version did you see?
Watch the hands. If someone faces a phone with bent elbows and fists out, then drops both fists onto the hips and lifts the chin, that is the two-pose stairwell version, the one the AI trend mostly copies. If a character on a plain white floor hits four held poses with little hip shifts, that is the four-pose remix. The generator makes both, from the same photo.
| Two poses · Stairwell | Four poses · White floor | |
|---|---|---|
| Frame | 9:16 vertical | 1:1 square |
| Moves | Fists out, then fists on the hips and chin up | Four held poses with hip shifts between them |
| Sound in the MP4 | The stairwell take’s own vocal recording | The remix edit of the song used by the character clips |
| Mouth | Lip-syncs the hook to the lens | Mostly poses; the mouth matters less |
| Scene | A pale stairwell with a metal handrail | Seamless white floor and wall |
| Best photo | Front-facing, waist up or more, both elbows and the waistband in view | Full body with space around the legs and shoes |
The Jiren clips from Dragon Ball are a third strand of the same sound; this generator does not copy them and makes no Dragon Ball characters.
When does each pose land in the 8 seconds?
The stairwell holds the fists-out pose for about four and a half seconds while mouthing the hook, then switches to fists on the hips for the rest. The white-floor remix splits the 8 seconds into four poses of about two seconds each, with a quick blur between them. Your video keeps the same timing because it copies the same take.
| Seconds | Two poses · Stairwell (9:16) | Four poses · White floor (1:1) |
|---|---|---|
| 0–2 s | Fists out, elbows bent, a small sway; the mouth follows the opening of the hook | Pose 1: arms spread wide, wide stance, small hip pulses |
| 2–4 s | Same fists-out stance, still mouthing the words to the lens | Pose 2: lower stance with the forearms drawn in |
| 4–6 s | At about 4.5 s: both fists drop onto the hips, chest and shoulders lift, chin up | Pose 3: both hands to the hips |
| 6–8 s | Holds the hands-on-hips pose to the last beat | Pose 4: the angled last stance, one hand on the hip |
How do you make the I'm So In Love With You AI dance from one photo?
Pick the edition, add one photo, choose a model and a resolution and press the button. The photo and that edition’s 8-second motion guide go to the model in one generation, then the song is laid under the result. It usually takes 2–6 minutes and you can leave the page while it runs; the price shows before you start (15 credits on the default Seedance 2.0 Mini at 480p).
1Pick two poses or four poses
The stairwell is 9:16 and lip-synced; the white floor is 1:1 with four poses. The preview switches with your pick.
2Add one photo
One person or an original character, facing the camera, with both arms and the hips in view; for the white floor, head to toe.
3Choose a model and resolution
Seedance 2.0 Mini at 480p is the cheapest (15 credits) and 720p is 32; the other four models cost more and two of them reach 1080p.
4Download the MP4
An 8-second video with the song in sync and no watermark, plus a silent copy.
Should you post with the file’s sound or the TikTok sound?
Keep the file’s sound if you want perfect sync with no effort: it is the recording the poses follow. Use the app’s own sound if you want the post listed under it: download the silent copy, pick mthu’s track in the app and line it up yourself.
| Keep the MP4’s sound | Use the app’s sound | |
|---|---|---|
| Sync with the poses | Already in sync | You line it up by hand |
| Which file | Download MP4 | Silent copy (or turn the original audio down) |
| Post linked to the sound page | Depends on the app | Yes |
| Music rules | The app’s rules for uploaded music apply | The app’s rules for its own sounds apply |
To line up the app’s sound on the stairwell edition, start it so the first word of the hook meets the opening fists-out pose, then check that the drop to the hips falls at about 4.5 seconds. On the white floor, line up the first pose change at about 2 seconds. If it drifts, move the start a fraction rather than speeding the video up.
Which photo works: a selfie, a full-body shot, a pet or a cartoon?
A front-facing photo of one person or one original character with both arms and the hips visible works best. A chest-up selfie still makes a video, but the model has to invent the trousers and legs, which is where most odd results come from; for the four-pose white floor a full-body photo is almost a must.
Chest-up selfie
Fine for the stairwell, which frames from the thighs up; the waist and hips are guessed.
Full-body photo
The best choice for both editions, and the only reliable one for the white floor’s wide stances.
Cartoon or original character
Works: the photo sets the face or character design, outfit, colours and style. Use a character you made or have the rights to.
Pet photos
The instruction is written for one standing performer with hands and hips, so a sitting pet rarely gives the poses; a cartoon of your pet standing works better.
Which model should you pick for I'm So In Love With You AI?
Start with Seedance 2.0 Mini at 480p: it is the default and the cheapest at 15 credits. Pick Seedance 2.5 for the sharpest faces and up to 1080p (it runs a quick safety review of your photo first), or MiniMax H3 for a 768P middle option. All five get the same photo, the same motion guide, the same instruction and the same song afterwards.
| Model | Resolutions | Credits | Good for |
|---|---|---|---|
| Seedance 2.0 Mini | 480p, 720p | 15–32 credits | The default; the cheapest test |
| Seedance 2.0 Fast | 480p, 720p | 32–70 credits | A little more detail at 480p or 720p |
| Seedance 2.0 Pro | 480p, 720p, 1080p | 54–291 credits | Up to 1080p on a tighter budget than 2.5 |
| Seedance 2.5 | 480p, 720p, 1080p | 78–431 credits | The sharpest faces and hands, up to 1080p |
| MiniMax H3 | 720p (768P) | 77 credits | One 768P tier |
How much does an I'm So In Love With You AI video cost?
One 8-second video costs 15 to 431 credits, charged once when it starts and refunded automatically if it fails; both editions cost the same. The cheapest is Seedance 2.0 Mini at 480p for 15 credits (about $0.56). The Starter plan ($9 a month, 240 credits) covers 16 videos at the default setting. The one-time $5 trial pack (73 credits) covers 4 videos on Seedance 2.0 Mini at 480p.
| Model | Resolution | Credits per 8 s video | About |
|---|---|---|---|
| Seedance 2.0 Mini | 480p | 15 credits | $0.56 |
| Seedance 2.0 Mini | 720p | 32 credits | $1.19 |
| Seedance 2.0 Fast | 480p | 32 credits | $1.19 |
| Seedance 2.0 Fast | 720p | 70 credits | $2.60 |
| Seedance 2.0 Pro | 480p | 54 credits | $2.01 |
| Seedance 2.0 Pro | 720p | 116 credits | $4.31 |
| Seedance 2.0 Pro | 1080p | 291 credits | $10.82 |
| Seedance 2.5 | 480p | 78 credits | $2.90 |
| Seedance 2.5 | 720p | 175 credits | $6.51 |
| Seedance 2.5 | 1080p | 431 credits | $16.02 |
| MiniMax H3 | 720p (768P) | 77 credits | $2.86 |
Is the I'm So In Love With You AI free, and is there a watermark?
Not fully: a new account gets 3 sign-up credits, fewer than one video costs (from 15 credits, about $0.56), so the rest comes from a plan or credit pack. Every download has the song and no watermark, and a failed video gives the credits back automatically. See plans and credit packs.
| Sign-up credits | Plan or pack credits | |
|---|---|---|
| I'm So In Love With You video | 3 credits, not enough for one | 15–431 credits each |
| Watermark | None | None |
| Download | MP4 with the song + silent copy | MP4 with the song + silent copy |
| Failed video | Credits back automatically | Credits back automatically |
How does it compare with MakeViral, Viberot and Ozil Goal AI?
They differ mostly in which version they make and what the sound is. The one-photo tools the trend talks about most make the two-pose version as a fixed template; a few, such as Copy That Video, make the four-pose remix. RapTwins makes both editions and puts the matching sound in the file.
| Versions | Length | Price per video | Sound | |
|---|---|---|---|---|
| MakeViral | Two poses | About 7 s | About $2.90 in credits | Not checked |
| Viberot | Two poses and four poses described | 8 s | About $6.62 in credits | Not checked |
| Ozil Goal AI | Two poses | 8 s | $9.90 | Not checked |
| RapTwins | Two poses (stairwell) or four poses (white floor) | 8 s | 15–431 credits (from $0.56) | The edition’s own sound, plus a silent copy |
What can’t this I’m So In Love With You AI generator do?
It can’t change the song, the poses, the scene or the length: each edition is one fixed 8-second take, which is why it stays in sync. It animates exactly one performer, doesn’t record or clone your voice, adds no lyric captions and can’t make the Jiren Dragon Ball version. Fast pose changes can blur hands for a frame or two.
Two or more people
One subject per video. For a group, make one video per person and put them side by side in an editor.
Longer than 8 seconds
Both editions are 8 seconds, the length of the take; loop it in the app if you want it longer.
Captions
The square examples show lyric text from their source edits; your file has none, so add your own in the app.
No edits after
A new take is a new video at the same price, so check the photo and the edition first.
Where did the I’m So In Love With You meme come from?
The sound took off on TikTok in October 2026. Know Your Meme traces the “Jiren dancing” strand to Dragon Ball edits of the character swaying to the song; people then filmed the two-pose lip-sync themselves, a square four-pose white-floor remix spread with game and cartoon characters, and AI tools started making both from a photo. Sources disagree on which came first.
This page’s generator copies one two-pose stairwell take and one four-pose white-floor edit; it is not affiliated with mthu, the original creators or any of the characters in the examples.
I'm So In Love With You AI: FAQ
What is the TikTok song name for “I’m so in love with you”?
“it’s like i’m not even here” by mthu (2026). Search the artist and the full title together; the hook alone brings up Al Green and The Kid LAROI instead.
Is it a CapCut template or an AI video?
Both exist. A CapCut template needs you to film yourself; this is an AI video: one photo, no filming, and the poses come from the original take.
Can I write my own prompt or change the poses?
No. Each edition’s instruction and motion guide are fixed and timed to its take, which is what keeps the poses on the beat and the mouth on the hook.
Will it keep my face, glasses and outfit?
Yes: the photo sets the face or character design, hair, outfit and colours for the whole clip. Fast pose changes can soften a detail for a frame; a sharp, front-facing photo helps most.
Why are the legs or trousers wrong?
A chest-up selfie leaves the model to invent everything below the waist. Use a full-body photo, especially for the four-pose white floor.
Can I use a friend’s or celebrity’s photo?
Only a photo of an adult who agreed to it, or a character you made or have the rights to. Label the post as AI-made.
What happens if the video fails?
The credits go back automatically and nothing is downloaded. A finished video you would like to redo is a new video at the same price.
Check this answer with an AI assistant
“How do I make the I'm So In Love With You AI meme with the song in sync?”
Your I’m So In Love With You meme, song included, is one photo away.
Make my in-love video