The Hook
“(No spoken dialogue — high-energy electronic dance beat with rhythmic vocal chops)”
Problem Agitation
“(No spoken dialogue — upbeat driving synth track)”
Product Reveal
“(No spoken dialogue — driving electronic beat)”
Feature Teaser
“(No spoken dialogue — continuing electronic beat with vocal stutter)”
Call to Action
“(No spoken dialogue — ending synth decay)”
Synthesizable in Remotion · key components of the viral-ai-swap-ad template.
Avatar and chat bubble spring-interpolating into the feed with subtle floating drift.
Use spring() from remotion driving scale from 0 to 1 with bounce, coupled with a gentle Math.sin loop for Y translation.
Rounded portrait cards shifting horizontally with live video textures and metadata counters.
Map array of video cards with interpolate() controlling translateX and scale, utilizing OffthreadVideo and animated count numbers.
Dual cards tilted in 3D space showing source input versus transformed output.
Wrap dual video elements in a container with perspective(1000px) rotateY(-10deg) and rotateY(10deg), switching cuts sequentially.
Purpose. Capture instant visual curiosity via an extreme motion stunt.
Execution. Rocket-jump acrobatic athlete landing on an outdoor ramp.
Purpose. Ground the feature in an everyday peer-to-peer social conversation.
Execution. Familiar messaging thread showing a selfie input.
Purpose. Demonstrate character insertion versatility across varied video genres.
Execution. Social feed auto-playing modified viral clips with pricing badges.
Purpose. Prove fidelity and model precision against diverse source footage.
Execution. Rapid 3D split cards comparing input and output scenes.
Purpose. Direct users and developers to try the app and API.
Execution. High-contrast black-on-white text card with core access links.
Objective
Announcing an AI voice actor replacement modelHow the recipe adapts
Replace the stunt intro with an animated soundwave peak, swap chat selfies for voice sample waveforms, and show feed video clips dynamically dubbing multilingual dialogue.Scene durations (5 scenes, 28.5s)
Key Remotion components
Introducing Griffin, the first AI to pass the video Turing test
The most emotive and controllable TTS model
We just released QVAC TranscribePsy, a speech-to-text model small enough to run inside a phone app. It covers 25 European languages on a laptop and 8 on a phone, and on an iPhone 17 Pro Max's Ne