OpenAI Developers · Launch Video Breakdown: Hook, Pacing & Motion Design
Shape a voice with a prompt or pick international presets
Scene-by-scene timeline & spoken transcript
The Hook
Sarcastic AI Intro
“Oh wonderful, another brilliant voice demo.”
- On screen
- GPT-Live Penny Ada Marvin
- Camera
- Static over-the-shoulder medium close-up on laptop screen.
- Motion
- Fluid audio-reactive mesh orb visualization.
Product Reveal
Introducing Voice Creation
“We're expanding voice options in GPT-Live 1 so you can find or create the right voice for your app.”
- On screen
- Voice Creation with GPT-Live 1
- Camera
- Medium shot of host sitting at a wooden table with laptop and mug.
- Motion
- Clean title card overlay with subtle fade.
Feature Teaser
Prompt-to-Voice Overview
“With prompt-to-voice, you can generate a voice from a high-level prompt, but also shape finer details like pitch, timbre, texture, and more. I'm working on a few new prototypes. Can you help me create some voices for them?”
- On screen
- (No on-screen text)
- Camera
- Close-up cut on speaker, then cut to medium profile shot.
- Motion
- Direct camera address with natural pacing.
Feature Teaser
Creating Penny (Shopping Assistant)
“Sure thing. Just tell me what vibe or character you're imagining and I'll spin it up. First up is a shopping assistant... Here's Penny, bright, quick, and curious. Hi, I'm Penny! Who are we shopping for? Tell me about them.”
- On screen
- GPT-Live Penny
- Camera
- Tight shot of laptop display, tracking generative orb animations.
- Motion
- Dynamic morphing fluid orb and audio-synchronized ripple waveform.
Feature Teaser
Creating Ada (Tutor Persona)
“I'm also working on an education app... Here's Ada: warm, thoughtful, and quietly playful. Hello, I'm Ada. What are we trying to understand?”
- On screen
- GPT-Live Penny Ada
- Camera
- Alternating medium presenter shot and screen close-up.
- Motion
- Sequential orb spawn with smooth horizontal layout reflow.
Feature Teaser
Creating Marvin (Deadpan NPC)
“Last one, I'm building a video game... deeply unimpressed with everything... Here's Marvin: gloriously gloomy and completely unimpressed. Oh wonderful, another brilliant voice demo. Perfect Marvin, you're hired.”
- On screen
- GPT-Live Penny Ada Marvin
- Camera
- Cut between host reactions and screen orb focus.
- Motion
- Pulsing audio reactive animation on purple orb.
Call to Action
Developer CTA & Sign-off
“The voices you create are saved to your account so you can use them in your own GPT-Live projects. Try prompt-to-voice in the Playground or via the API today. Happy building.”
- On screen
- (No on-screen text)
- Camera
- Static medium shot with host looking into lens.
- Motion
- Natural conversational fade-out.
Motion design primitives
Synthesizable in Remotion · key components of the ai-voice-persona-showcase template.
FluidVoiceOrb
Dynamic 3D gradient sphere deforming subtly with audio amplitude.
+ Code construction recipe
Use Three.js canvas in Remotion. Apply Simplex noise vertex shader modulated by RMS audio levels.
CharacterOrbRoster
Horizontal flex row expanding smoothly as new voice spheres spawn.
+ Code construction recipe
Spring-animated layout translation mapping array additions to interpolated horizontal positions.
Style DNA & aesthetic system
Visual aesthetic
Typography
Motion
Special effects
Pacing & rhythm
Conceptual story rhythm
- 1
Hook
Purpose. Disarm viewer with witty deadpan AI preview.
Execution. Close-up of laptop with live pulsing orb.
- 2
Problem / Solution
Purpose. Introduce prompt-to-voice capability.
Execution. Presenter speaking directly to camera.
- 3
Demonstration Arc
Purpose. Prove versatility across commerce, learning, and gaming.
Execution. Prompting live AI assistant with on-screen additions.
- 4
Call to Action
Purpose. Direct developers to API and Playground.
Execution. Warm wrap-up on presenter with laptop.
Remix in Animatiq Studio
Master remix prompt · agent-ready
Sample remix
Objective
Announcing real-time tone customization in a customer support AI platform.How the recipe adapts
Replace casual consumer archetypes with professional enterprise customer service voice personas.Remix blueprint · ai-voice-persona-showcase
Scene durations (5 scenes, 75s)
Key Remotion components
Related AI & ML Product Launches
Explore all AI & ML launches →
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