Two routes. The manual route is for people who want to write a still, then a motion pass, in tools they already pay for. The one-tap route is the form on the studio and on the generator page. This page prints the manual wording and then says what TrendShot runs instead.
Do not paste a commercial recording into a prompt and ask a model to copy it. Add sound in an editor if you are on the manual route. TrendShot muxes the demo soundtrack at download and does not send a song in the generate request.
What you need before a manual pass
You need one photo of the face you want to keep, a video tool that accepts an image and a motion description, and an editor if you want a specific track. You also need to pick a version. The meadow hop, the glass still, and the onesie are three pictures. Mixing all three sentences into one prompt is how you get a window in a field.
For the meadow, a pet photo should show the face large, both eyes open, ears included. A human headshot should be chest-up, no sunglasses. Say in the prompt that the body and the hop come from the reference clip if your tool has a reference slot. If it does not, describe the hop in words and accept that the timing will wander.
Budget a retry. Paws smear. Faces drift back to the reference animal. A second animal appears when the photo had two. None of that is fixed by adding more adjectives about how cute the eyes are.
The scene you are matching
Use the credited meadow clip as the picture in your head, in this order. We measured that file at 16.7 seconds. A manual 5-second pass will only cover the start of this table. A TrendShot export is about ten seconds and covers the approach plus the raised paws, not the final seconds of the long cut.
Shot table for the meadow file
These rows are the hosted file, not a guess at every fan edit. If your manual tool cannot hold a camera this steady, you will get a different meme.
| Beat | What the frame holds | What should not change |
|---|---|---|
| Approach | Grass, face, paws coming forward, petals in the sky | Do not cut to a new room |
| Pads | Both front paws up, pads toward the lens | Do not add a third paw |
| Tilt | Head tips, one paw nearer the camera | Do not swap in a second animal |
| Fill | Face larger, paws still up | Do not burn caption text into the frame |
Prompt: pet face on the meadow hop
Upload the pet photo. Tell the tool: keep this animal's face, ears, muzzle, and face markings. Put that face on a fluffy snow-leopard body with pale fur and dark rosettes. The setting is a green meadow under a blue sky with a few pink petals. The animal hops toward the lens and holds both front paws up, pads forward, mouth closed, eyes wide and calm. One animal. Two front paws. Vertical frame. No text, no logo, no glass window.
Then a motion line: one continuous shot, camera almost still, slight sway. The body comes closer. The paws stay toward the lens. Only the head, ears, and front legs move. Do not ask for a song by name. If the tool makes audio, ask for light outdoor air and soft movement, then replace the track in an editor if you need a specific sound.
Regenerate if the face becomes the generic leopard from the reference, if a paw splits into extra toes, or if the background turns into a zoo window. Those are the three misses this prompt is meant to prevent.
What the motion reference controls
If your tool accepts a reference video, that clip should own the hop, the petal drift, the grass, and the camera. The photo should own the face. Write that split in the prompt so the tool does not average them into a third animal.
On TrendShot this split is already written. Video 1 is the muted meadow file. Image 1 is your upload, sent more than once so the face outweighs the clip. You do not paste the paragraph above into our form. The form has no prompt box.
A reference clip with sound attached is the wrong file to send a provider if the goal is to avoid copying a recording. Our API file is muted. Sound is muxed later from the demo track, at download.
What the photo controls
The photo controls identity: eyes, muzzle or nose, mouth, ears, face markings, and, for a human headshot, the person's face. It does not control the meadow. It does not control the number of paws. It does not add a caption.
A full-body human photo also controls clothes and body shape. That is a product rule, not a prompt flourish. If you are writing a manual prompt and you want the spotted coat, say the coat comes from the snow-leopard reference and the face comes from the photo. If you upload a full-body portrait to TrendShot, the coat can lose, because full appearance includes clothing.
Pets are a face replace onto the clip's body. That is allowed here even though the body is not human. Do not write a line that keeps the video animal's face and only recolors the fur. That line fights the point of uploading your pet.
What to regenerate
Regenerate when the face in the result is still the meadow leopard. The usual cause is a small or soft face in the upload. Crop closer and run again.
Regenerate when you see a window, snowy rocks, or a hooded onesie and you asked for the meadow. The prompt picked up a different version of the meme. Delete those words and name the grass and the petals again.
Regenerate when limbs multiply. Say one animal and two front paws in the prompt. Fast hops still smear pads. A second try is the practical fix. There is no slider that guarantees toe count.
The glass-window version, if you insist on writing it
Some posts stay on the 2024 pose: the animal behind glass, both paws flat, a snowy enclosure soft behind it, then one paw taps the glass. That is a real branch of the meme. It is not the clip this site generates. If you write it by hand, say the photo is seen through a window, both paws on the glass, then one paw taps twice and returns. Keep one animal. Do not also ask for a meadow in the same paragraph.
A human branch puts the person in a spotted hooded suit with round ears and paw mittens, same glass pose, face unchanged. Say the clothes become the suit and the face stays the photo. Do not ask TrendShot's full-body mode to do this. Our full-body mode keeps the clothes in the photo.
A third branch adds a leopard at your shoulder and leaves your photo otherwise intact. That needs a prompt that says exactly one person and exactly one animal, the paw never covering either face. This site has one photo slot and will not run that scene.
Sound without copying a recording
For a manual edit, export the picture silent or with whatever neutral audio the tool made, then lay a track in an editor you have rights to use. Do not write the name of a commercial song into the prompt and ask for a copy.
The meadow clip's own audio can be replaced the same way. We did not publish a song title for it because we did not identify one. If you know the track and the app has it, add it there.
TrendShot's download already contains the demo soundtrack. You can still replace it. The generate prompt only asks for a fresh soundtrack so the provider call stays clear of a copied recording.
The one-tap alternative
Open the studio or the generator page. Upload one photo. Buy Single for $8.99 or Party Pack for $21.99 if the button is locked. Generate. Download a with-audio file of about ten seconds in 9:16. 30 credits per success. Failed jobs return credits. The pack lasts about 30 days on this browser.
You do not paste the meadow paragraph, the glass paragraph, or the onesie paragraph. The locked text generates a new clip from video 1 and takes the face from image 1. The glass and onesie paragraphs exist on this page so a manual editor can see them. They are not the product.
Credit the meadow performance to @noperscan if you show that reference. Your generate is a new clip with your face on that motion. It is not a reupload of the TikTok file.
Step-by-step for this trend: how to make the Snow Leopard Paws at You AI video.
Prompt and pipeline: Snow Leopard Paws at You AI prompt. Generate: Snow Leopard Paws at You AI video.
FAQ
Do I have to use these prompts on TrendShot?
No. The form does not take a prompt. These blocks are the manual route. The one-tap path is the studio.
Which prompt matches the button?
The meadow hop. Face from your photo, body and motion from the credited clip. Not the glass window and not the onesie.
Can I name a song in the prompt?
Do not ask a model to copy a commercial recording. Add sound in an editor, or use TrendShot's demo soundtrack and replace it later if you want.
What does the one-tap path cost?
30 credits. Single is $8.99. Party Pack is $21.99. About ten seconds, with-audio 9:16. Failed jobs return credits.