Kitten Knight vs Dragon: The AI Video Prompt
A free kitten knight AI video prompt for Seedance 2.5. A tiny ginger kitten in armor faces a giant dragon, screams, and runs into the camera. 16 seconds.
Have it running in three steps
- 1
- 2Open ImagineArt
- 3Paste it and generate
What you'll need
30 min- ImagineArt
This kitten knight AI video prompt makes a 16-second dark fantasy clip with a funny ending. A tiny ginger kitten in silver armor stands on an empty battlefield at dusk. It lifts its sword like a real hero.
Then a giant dragon rises out of the fog and roars right over its head. The kitten loses all its courage. It drops the sword, screams and runs straight at the camera. It crashes into the lens, the camera falls, and the screen goes black.
You need two pictures: the kitten and the dragon. This page gives you a prompt for each one, and the video prompt that puts them together. No photo of you is needed.
The whole process
- Make the kitten with the first image prompt below.
- Make the dragon with the second image prompt.
- Open ImagineArt, go to video and pick Seedance 2.5.
- Add the kitten as @Image1 and the dragon as @Image2.
- Set 16 seconds and 9:16. Paste the video prompt and generate.
What happens in 16 seconds
| Time | What happens |
|---|---|
| 0–3s | The kitten stands in the mud with its sword, broken spears all around |
| 3–5s | Side view. It lifts the sword above its head, brave and serious |
| 5–9s | Wide shot from behind. The dragon rises out of the fog and roars over the tiny kitten |
| 9–11s | Close-up. The kitten stares up with huge eyes, mouth open |
| 11–14s | The sword lies in the mud. The kitten turns and runs at the camera, screaming |
| 14–16s | It hits the lens. The camera flips up to the sky, then black |
Make the kitten
Open an image model in ImagineArt+200 free credits. We used Nano Banana Pro. Set the shape to 9:16 and paste this prompt. It needs no reference image.
Ultra photorealistic full-body character reference of a tiny ginger kitten, about 8 weeks old, standing upright on its hind legs like a little knight. Round fluffy face, soft orange and cream tabby fur, huge expressive amber eyes, small pink nose, real kitten proportions. It wears polished silver medieval plate armor fitted to its tiny body: engraved breastplate, pauldrons, chainmail at the waist and legs, small armored paws, and a short weathered deep red cape. A small knight’s sword hangs in a scabbard at its waist, both front paws resting on the hilt. Brave but slightly nervous expression, looking up past the camera. Setting: misty medieval battlefield at dusk, wet dark grass and mud, broken spears stuck in the ground, cold blue-grey cloudy sky with a faint warm glow on the horizon. Low camera at the kitten’s eye level, shallow depth of field, cinematic soft light, realistic fur detail, realistic metal reflections, live-action fantasy film still. No cartoon, no 3D render, no anthropomorphic human body, no text, no logos.
This is the kitten we got. The prompt asks for the sword in a scabbard, but the model put it in the kitten’s paw. That is fine, because the sword has to be in its paws in the video anyway.
Make the dragon
Make a new image at 3:2. Our dragon is our own design. We made this picture from a drawing of it, so our prompt started with a line about the reference. Below is the same prompt with that line replaced by a description in words, so you can use it without a drawing.
If you have your own dragon picture, add it as a reference. Then start the prompt with: “Use the dragon from the reference image only for its anatomy, shape and colors. Do not copy its drawing style.”
A huge wyvern with two hind legs, wings that work as front limbs, torn wing membranes, a spiked crest, a long spiked tail, dark red and black scales, amber eyes with slit pupils.
Show it as a real living animal photographed on location, like a frame from a live-action film shot on ARRI Alexa with an anamorphic lens. Three-quarter view from low on the ground, the dragon standing on a muddy misty battlefield at dusk, head slightly lowered and turned, not facing the camera straight on, wings half open and resting on the wing joints.
Realism details: matte, weathered scales with dirt, dried mud, old scars and chipped edges, scales of uneven size, dull desaturated dark red, not glossy. Wet skin on the wing membranes, light passing through the thin torn parts. The dragon is far away and huge, so fog and atmospheric haze soften its lower body and lower the contrast with distance. Backlit by a cold blue-grey overcast sky with a faint warm glow on the horizon, rim light on the crest and wing edges, the front of the body falls into soft shadow. Breath steaming from the nostrils. Fine film grain, natural soft focus, no sharpening.
Not a 3D render, not a video game, not concept art, no symmetric pose, no saturated glossy red, no text.
Check the dragon before you move on. If it looks shiny, smooth or like a toy, make it again. The video will copy that look.
Set it up and generate
Open ImagineArt+200 free credits, go to video and pick Seedance 2.5.
Add the kitten first, so it becomes @Image1. Add the dragon second, so it becomes @Image2. Set the length to 16 seconds and the shape to 9:16. We used 720p. Paste the prompt below as it is.
[LOOK — LOCKED, APPLIES TO EVERY SHOT] ARRI Alexa LF with Panavision C-series anamorphic lenses, T2.0: oval bokeh, soft frame edges, a faint horizontal flare only on the warm horizon light, shallow focus that drops the background out one step behind the subject. Fine organic grain, soft highlight roll-off, deep but detailed shadows, medium contrast, unsharpened natural texture. Cold Dusk grade: steel blue-grey sky and fog, desaturated olive grass, warm orange only on the horizon glow and the dragon’s eyes. Real handheld camera operation with weight and slight drift, natural motion blur on movement, no perfectly still frames. Live-action fantasy film, no music.
[REFERENCES] The kitten corresponds to @Image1. Use only the kitten’s face, ginger tabby fur, eye color, body size, silver engraved armor, chainmail, deep red cape and the sword. Do not use the background, the spears, the shields or the composition of @Image1. The dragon corresponds to @Image2. Use the dragon’s anatomy, matte weathered dark red and black scales, torn wing membranes, spiked crest and amber eyes, and keep the same photographic realism: dirt and scars on the scales, fog softening its body, backlight from the sky. It is a wyvern: two hind legs, the wings work as front limbs. Do not use the composition of @Image2. There is exactly one kitten and one dragon in the whole video.
[GENERATION GOAL] A 16-second dark fantasy scene with a comic twist. A tiny ginger kitten knight bravely raises its sword on an empty battlefield. A giant dragon rises out of the fog and roars at it. The kitten sees it, drops the sword, turns around, screams, runs straight at the camera and crashes into it. The camera falls and the screen goes black.
[STAGE 1 — 0–3 sec] Initial state: the kitten is a few meters from the camera on a misty battlefield, sword in its right paw, tip pointing down. Wet dark grass, mud puddles, broken spears and a fallen shield around it. Primary event: the kitten walks toward the camera on its hind legs with small, careful, clearly visible steps, the sword tip dragging a thin line through the mud. Camera: low-mode Steadicam about 15 cm above the mud, 40mm lens, moving backward in front of the kitten at its walking speed. Mud splashes under its paws, the red cape sways, the fur is damp at the tips. End state: the kitten stops, center of frame, ears forward.
[STAGE 2 — 3–5 sec] Continue from the previous stage: same kitten, same armor, sword in the right paw. Primary event: the kitten raises the sword above its head with both front paws. The blade catches a cold silver glint. Its whiskers push forward, its mouth closes in a brave, serious look. Camera: cut to a medium shot from the kitten’s left side at a low angle, 50mm, slow push-in toward its face. End state: the kitten holds the sword raised, looking up.
[STAGE 3 — 5–9 sec] Continue from the previous stage: the kitten holds the sword raised. Primary event: the dragon from @Image2 rises slowly out of thick fog in front of the kitten, lowers its head close to it and roars. The blast of hot breath flattens the kitten’s fur and whips its cape backward. The dragon is far larger than the kitten and partly hidden in fog; its body is darker and softer with distance. Camera: hard cut to a wide shot from low behind the kitten and to the right, 24mm. The kitten is a tiny figure at the bottom of the frame. The dragon’s head is turned to the side, seen in three-quarter profile with one eye toward the camera, never straight on and never symmetrical. The handheld frame shakes slightly with the roar. End state: the dragon’s open jaws are a few meters above the kitten, the kitten still stands with the sword raised.
[STAGE 4 — 9–11 sec] Continue from the previous stage: the dragon is now off-screen. Primary event: the kitten realizes what it is facing. Its eyes go wide, its ears flatten, its fur puffs up. Its paws open and the sword falls out of them, clearly visible as it drops and splashes into the mud at its feet. Camera: cut to a close-up of the kitten’s face, low, 50mm, the kitten looking up out of the frame. The dragon is not visible in this shot. Behind the kitten there is only thick grey fog and the empty battlefield, out of focus. End state: the kitten stands without the sword, the sword lies in the mud, the kitten is frozen with wide eyes.
[STAGE 5 — 11–14 sec] Continue from the previous stage: the sword stays in the mud, the dragon stays off-screen. Primary event: the kitten spins around toward the camera, opens its mouth wide in a scream and runs straight at the camera on all four paws, cape flapping, armor bouncing. Camera: a low handheld shot at ground level in front of the kitten, 35mm, held in place. The kitten grows larger in frame with every stride. The dragon is not visible in this shot: behind the kitten there is only thick grey fog, broken spears and the empty battlefield. End state: the kitten’s screaming face fills almost the whole frame.
[STAGE 6 — 14–16 sec] Primary event: at about 14 seconds the kitten leaps and crashes into the lens with its armored chest. The camera is knocked over: the frame jolts, spins and falls sideways into the mud. Camera: the image tilts and drops to the ground, then cuts to a completely black screen. End state: a fully black screen until the end.
[AUDIO] Stage 1: <wind across the field> <soft wet squelch of small paws in mud> <light clink of armor plates>. Stage 2: <metal ring of the sword being raised> <cape flapping in the wind>. Stage 3: <heavy footsteps that shake the ground> <a deep, loud dragon roar> <rushing wind of its breath>. Stage 4: <the sword splashing into the mud> <a small frightened gasp>. Stage 5: <a long, high, panicked kitten yowl> <fast rattling of armor plates> <paws slapping the mud>. Stage 6: <a hard thump against the camera> <the camera clattering into the mud> then silence on the black screen. No music, no dialogue.
[MAINTAIN CONSISTENCY] Keep the kitten’s face, ginger fur, size, armor and red cape the same in every shot. The sword is in the kitten’s paws until it drops it in Stage 4. The kitten is a real kitten with real kitten anatomy, not a person in a costume. The dragon keeps two hind legs and wings as front limbs, with no extra arms. The dragon appears only in Stage 3. One kitten, one dragon, no text on screen, no logos, no subtitles.
Why this prompt works
The first block sets one camera and one color look for the whole video. Seedance 2.5 makes several shots in one go. Without this block, each shot can come out with a different grain, sharpness and color, and the cuts stop looking like one film.
Each camera name comes with what it should look like on screen, such as soft oval blur in the background and fine grain. The model copies the look you describe, so a camera name on its own does little.
Each image has one job, and the prompt says what not to take from it. Without that, the shields and spears from the kitten picture would show up again in every shot.
The video is split into six stages. Each stage has one main action and says how the shot should end. This keeps the model from mixing up the order, so the roar comes before the panic and the crash comes last.
The dragon appears in one stage only. The prompt says twice that it is not visible in the close-up and the run. In an earlier version without those lines, the dragon stayed behind the kitten the whole time and the run felt less funny.
Every stage has its own sounds: the roar, the sword falling in the mud, the kitten’s scream, the thump on the camera. The scream is what makes the ending land, so keep the sound on when you post it.
Where it goes wrong
| What you got | The fix |
|---|---|
| The dragon looks like a 3D game render | Your dragon picture is too clean or drawn. Make it again with the second prompt and keep “Not a 3D render, not a video game, not concept art” |
| The dragon is still behind the kitten when it runs | Keep “The dragon is not visible in this shot” in Stages 4 and 5 |
| The dragon’s face looks straight into the camera | This happened in all our tries, even with “never straight on”. It still works. A dirty, foggy dragon picture makes it look less fake |
| The kitten barely walks at the start | This happened in our tries too. The first three seconds are short, so it does not hurt the video |
| The sword vanishes instead of falling | Keep “clearly visible as it drops and splashes into the mud”. The model sometimes skips it, so generate again if you need that moment |
| The kitten looks like a person in a cat costume | Keep “not a person in a costume” and “real kitten anatomy”. Check the kitten picture stands like a kitten |
| The wrong picture plays the wrong part | Add the kitten first and the dragon second. The first picture is @Image1 |
| ImagineArt says “Asset already exists” | This can happen when you reuse a picture from an earlier video. Upload the picture again and use the new copy |
Questions people ask
Which AI model makes this video?
We made it with Seedance 2.5 on ImagineArt+200 free credits. The video prompt is written for it: 16 seconds, 9:16, two reference images.
Can I use my own cat?
We have not tried it, but it should work. Add a photo of your cat as a reference when you make the kitten picture. Start the kitten prompt with “Use the cat from the reference photo. Keep its face, fur color and markings.” Then change “ginger” in the video prompt to your cat’s color.
Can I use a different dragon?
Yes. Use any dragon picture as @Image2, as long as it looks like a photo. If your dragon has four legs, delete the line “It is a wyvern: two hind legs, the wings work as front limbs” and the matching words in the last block. We only tested our two-legged dragon.
Is it vertical?
Yes, 9:16, ready for Reels, TikTok and Shorts.
Is it free?
The prompts are free. ImagineArt runs on credits, so the pictures and the video use credits from your account.
Want more scenes like this? See all our video prompts, or open ImagineArt+200 free credits and make your kitten.
Why this workflow runs here
Everything in one place
Free to start
Minutes, not days
Make yours in the next 30 min
You have the whole prompt and every setting. The only thing left is an account, and that takes a minute.
- 200 free credits on sign-up
- Prompts ready to paste
- Upgrade only if you want more
Referral link — if you subscribe I earn a commission, at no extra cost to you. More on that.
Related prompts
Cinematic Free Action Scene AI Video Prompt: Bike Jump and Fight
Action scene AI video prompt for Seedance 2.5. You jump a futuristic bike off a ramp, land in fire and fight on a burning highway. 15 seconds, one photo.
Cinematic Free Anime Borscht — The Whole Recipe in One 25-Second Video
One prompt cooks Ukrainian red borscht from raw meat to a served bowl, in glossy anime style. Six stages, hands and food only, vertical for Reels.
Cinematic Free APT AI Prompt — A Dog and a Cat Do the APT. Dance
Free APT AI prompt for Seedance 2.5. A dog sings and a cat plays drums in the APT. dance. Download the clip, make two sheets, swap in your own pets.