viralprompts
Now decoding Seedance 2.0 + Gemini Omni Flash

Stop guessing.
Steal the prompts behind viral videos.

A vault of the internet's most viral AI videos — with the exact prompts, models, and settings used to make them. Paste any clip and get its full recipe.

7-day free trialNo prompt-writing experience needed

INPUT / VIRAL CLIP ANALYSIS READY
Seedance 2.000:04.80 / 00:06.00
MASTER PROMPT · AS USED

Extreme macro. Transparent crystal pastry fractures under a polished blade, revealing slow liquid starlight...

1clip uploaded

3–12scenes decoded

1copy-ready prompt recipe

built for short-form speed
BUILT FOR THE MODELS CREATORS ACTUALLY USE

Seedance 2.0

Gemini Omni Flash

Veo 3.1

Kling 3.0

Runway Gen-4.5

LIVE FORMAT LIBRARY

The feed, organized into repeatable systems.

Explore the trend radar
Glass-cut ASMRSeedance 2.0

Glass-cut ASMR

Macro material reveal

Mascot mayhemKling 3.0

Mascot mayhem

Character continuity

Tiny rescue arcVeo 3.1

Tiny rescue arc

Emotional micro-story

The model fingerprintGemini Omni Flash

The model fingerprint

Ranked generation clues

Seedance 2.0

Glass-cut ASMR

Macro material reveal

Kling 3.0

Mascot mayhem

Character continuity

Veo 3.1

Tiny rescue arc

Emotional micro-story

Gemini Omni Flash

The model fingerprint

Ranked generation clues

THE UNFAIR SHORTCUT

Six seconds of chaos.
Every decision exposed.

The final video is only the surface. ViralPrompts reveals the repeatable system underneath it.

01

It watches the whole clip

Not just the opening frame. ViralPrompts tracks scene changes, camera movement, pacing, sound cues, and the final loop.

02

It finds the generation logic

Each shot is mapped to subject, action, lens, lighting, continuity, and the model most likely to recreate it.

03

You get the exact prompts

Copy one master prompt or work scene by scene with timing, references, generation settings, and the model that made it.

01 · THE CLIP

A viral video looks like magic.

On the feed you only ever see the finished six seconds.

02 · THE BREAKDOWN

VaultVision AI shatters it into scenes.

Cuts, start frames, camera moves, pacing, and sound — every decision isolated.

03 · THE RECIPE

Every frame becomes a prompt.

An image prompt and an animation prompt for each scene, matched to the model that can shoot it.

THIS IS WHAT VIRALPROMPTS DOES TO EVERY CLIP.Analyze your first video
PROMPTS BUILT FOR THE FEED

Decode any format.
Copy the prompts.

From deliberate slop to cinematic product work, the teardown adapts to the grammar of the clip—not a generic prompt template.

Impossible ASMR
Kling 3.0

Impossible ASMR

Material physics, macro lens logic, loop timing, and native SFX.

Brainrot characters
Seedance 2.0

Brainrot characters

Character continuity, escalation beats, dialogue, and camera energy.

Animal micro-drama
Veo 3.1

Animal micro-drama

Hook, emotional turn, rescue arc, reaction shot, and score direction.

WHO IT'S FOR

Built for people
who actually ship.

Four ways creators put the breakdown to work the same day they run it.

Ship a daily slop seriesFaceless channels

Ship a daily slop series

Tear down the format that is printing views, lock your recurring characters, and publish new episodes every day.

Series-ready prompts
Turn references into shot listsShort-form editors

Turn references into shot lists

Clients say “make it like this.” Drop the reference in and hand back scene-by-scene prompts, camera notes, and timing.

Scene manifests
Decode competitor creativeBrand & UGC teams

Decode competitor creative

See exactly why a rival AI spot holds attention — hook, pacing, sound — then rebuild it inside your own brand world.

Retention signals
Learn model-native promptingAI filmmakers

Learn model-native prompting

Every breakdown is a masterclass: image prompt versus animation prompt, continuity locks, and the right model per cut.

Model matching
Three video frames converging into a visual prompt fingerprintVISUAL SIGNATURE RESOLVED
MODEL MATCH, WITH CONTEXT

One clip. A ranked generation fingerprint.

Visual artifacts, motion cadence, audio behavior, and available provenance become an evidence-backed model recommendation—not a false claim of certainty.

STRONG MATCHSeedance 2.0
POSSIBLEVeo 3.1
LOW SIGNALRunway Gen-4.5
NOT A VAGUE SUMMARY

The prompts, the models, the settings.

ViralPrompts reverse-engineers the clip into working instructions. Every prompt is organized, timed, and matched to the AI model that can actually generate it.

  • Scene-by-scene prompts and timing
  • Camera, lens, motion, lighting, and sound
  • Character and object continuity notes
  • Ranked model match with confidence signals
Run a free teardown

crystal-croissant-final.mp4

6.0 SEC
SCENE 02 · THE REVEAL

Fracture the shell. Keep the universe intact.

Crystal croissant scene breakdown02.10 — 04.80
100mm macroMaterial revealCrystal crack
PROMPT BLOCK

Camera remains locked on the blade contact point. A controlled fracture travels left to right while the internal nebula rotates at one-quarter speed...

TikTok Instagram
VIRAL PROMPT LAB / SHORT-FORM VIDEO

From viral clip
to usable prompt.

Capture a breakout TikTok or Reel, isolate the hook, camera, motion, sound, and loop, then export the generation recipe while the format still matters.

Decode your first viral clip
01 FIND 1M+ SIGNAL02 BREAK DOWN HOOK03 COPY THE PROMPT
ONE MEMBERSHIP. EVERY PROMPT.

Make the next one
before the trend is over.

Built for creators, editors, and lean social teams who need answers now—not another blank prompt box.

MEMBERSHIP

$14.99/month

Every viral prompt in the vault, plus the analyzer that reverse-engineers any clip you feed it.

  • Full vault of viral prompts, updated weekly
  • Scene-by-scene breakdowns with camera and lighting logic
  • AI video analyzer — upload or paste any viral link
  • Model matches: Seedance, Veo, Kling, Runway and more
  • Or $79/year — save 56%
Start your 7-day free trialCancel anytime in one click. No charge until the trial ends.
QUESTIONS, ANSWERED

The fine print,
without the fog.

Does this only work with AI-generated videos?

No. You can upload any short-form clip. ViralPrompts separates observable filmmaking choices from likely AI-generation choices, so you still get a usable shot plan for live action, mixed media, or fully synthetic videos.

Can it identify the exact model used?

No detector can prove a model from pixels alone in every case. ViralPrompts combines provenance metadata when available with visual and audio evidence, then returns ranked model matches with an explicit confidence level.

Do I need to know prompt engineering?

No. The finished recipe is written to be pasted into the recommended generator. Advanced controls stay visible when you want to tune camera, motion, timing, references, or audio.

What do I get from one analysis?

Feed it an upload or a TikTok / Instagram link and you get the master prompt, a scene-by-scene timeline with image and animation prompts, the model match, camera and lighting notes, and generation settings — ready to paste into the recommended generator.

THE CLIP IS ALREADY TELLING YOU HOW IT WAS MADE.

ViralPrompts helps you hear it.

Analyze your first video

7-day free trialSteal the prompts behind the feed.

Analyze