AI Cat Video Prompt: Doorbell Camera Cat Musician (Sora 2)
The doorbell-camera cat did 10M+ views because the joke has three parts: an absurd performance, a fixed camera that makes it feel found rather than made, and a human who walks in and argues with the animal. This prompt writes all three in — the midnight porch set, the timed dialogue beats, and the hiss payoff — so the clip lands as one continuous take instead of three stitched ideas. Trumpet and piano variations are inside.
The 10M-view doorbell-camera format: a cat busking on the porch at midnight, the door opening, and the argument. Dialogue and sound design written in.
Create a 12-second vertical video in Sora 2, 9:16, fixed doorbell-camera angle, grainy security-camera look, muted night colours, slight fisheye lens. [00:00] An orange tabby cat stands on its hind legs on a house porch at midnight, strumming an acoustic guitar with its front paws under a warm porch light. The cat plays confidently, head tilted back, tail swaying. Sound: gentle acoustic guitar, crickets in the background. [00:05] The front door swings open. A woman in pyjamas appears in the doorway, hair messy, and says, exasperated: "What are you doing?! It's midnight!" Sound: the door creak, her voice, the guitar stopping mid-chord. [00:08] The woman takes the guitar from the cat. The cat stays on its hind legs, hisses, and meows angrily at her, tail puffed up. She stares at the cat, speechless. Sound: one sharp hiss, loud indignant meows, then silence except for the crickets. One fixed camera position throughout, no cuts, no camera movement. Realistic cat movement, paws on the strings, natural night lighting from the single porch bulb. No text on screen, no captions, no music beyond the guitar.
See what this prompt creates
The format is everywhere on TikTok: a security camera, a porch at midnight, a cat doing something no cat should do — playing guitar while the owner opens the door and loses an argument with it. The clips past 10M views are not the prettiest renders. They feel found: one camera that never moves, one ridiculous performance, one human reaction. This prompt is engineered for that shape, and it sits in our video collection next to the Sora 2 cinematic template.
How to use it
- Paste the prompt into Sora 2 as written, keeping the time stamps and quoted dialogue. The stamps are the joke's timing — removing them lets the model rush the argument.
- Generate two or three takes and judge the reaction, not the guitar: the hiss and the stare are the shareable second. A perfect cat with a flat reaction is the wrong take.
- Post it as one continuous clip with no caption burned in.
Why this prompt works
Four mechanisms do the work. The fixed camera: one doorbell angle, no cuts, no movement. Dialogue written as dialogue: her line in quotes at its own time stamp. The named sound design: guitar, crickets, door creak, hiss — naming the sounds stops the model filling silence with invented music. And the payoff beat: the hiss at 00:08, because every viral version ends on the cat winning the argument.
Two more midnight musicians
- Trumpet cat: keep the camera, porch, stamps and dialogue, and replace the instrument and sound with: "a grey tabby cat stands on its hind legs playing a brass trumpet, cheeks puffed, under the porch light" and "sound: one long bold trumpet note, crickets". The payoff becomes the cat lowering the trumpet slowly and staring at her, unimpressed.
- Piano cat: replace the scene with: "a black cat sits upright at a small keyboard on the porch, playing with both front paws, head bobbing" and the sound with "plinking piano notes, crickets". Keep her line and the hiss beat word for word
Failure fixes: cat refusing to stand upright — simplify to "sits upright at the guitar, front paws on the strings". Paws sliding off the strings — cut the strumming to "paws resting on the strings, one paw moving" and let the sound carry it. Dialogue garbling — shorten her line to "It's midnight!".
How to use this prompt
- Copy it — hit the copy button above to grab the raw prompt text.
- Fill the brackets — replace anything in [BRACKETS] with your own details.
- Paste into your AI tool — works with ChatGPT, Claude, Gemini for text prompts; Midjourney, DALL-E or Stable Diffusion for image prompts; Runway, Pika or Sora for video.
- Iterate — tweak one bracket at a time until the output is exactly right.
Frequently asked questions
Does Sora 2 actually speak the dialogue in this prompt?
Yes, Sora 2 handles short spoken lines. Keep them under six words and in quotes at their own time stamp — long speeches garble, short exclamations land.
Why does the camera have to stay fixed?
Two reasons: a locked doorbell angle is what makes the footage feel real, and it stops the model morphing the cat between shots. Movement is the enemy of this format.
Can I use my own cat's photo as the reference?
Sora 2 takes an opening image: generate or upload a still of your cat on a porch first, then use this prompt for the motion. The tabby in the example is the generic version.
My cat comes out looking like a toy. What fixes it?
Restate 'realistic cat movement, natural fur, real paws' and drop any cute or cartoon wording. The grainy security-camera look also hides a lot of rendering softness.
