
Introducing GPT-Realtime-2 in the API: our most intelligent voice model yet, bringing GPT-5-class reasoning to voice agents. Voice agents are now real-time collaborators that can listen, reason, and
The Hook
“Hey everyone, we're introducing new real-time audio models in the OpenAI API. In this demo, I'll show two of them: GPT Realtime Translate for live translations, and GPT Realtime 2 for voice agents that can follow instructions and take actions. Let's start with translations, 'cause that one feels so magical.”
Product Reveal
“Ce qui est très impressionnant, c'est que le modèle peut m'écouter et traduire en même temps que je parle. Il sait attendre juste le mot important, comme le verbe, pour commencer directement à traduire. Et le résultat, c'est une conversation beaucoup plus naturelle, exactement comme un dialogue entre deux personnes.”
Feature Teaser
“Merci beaucoup Dom. Isn't that amazing? The model can translate across 70 different languages in real time, really following the shape of every sentence. For the next demo, let's talk about GPT Realtime 2, our new model that brings intelligent reasoning to voice agents.”
Feature Teaser
“Hi there. Hi again. What's up? Yeah, I have a customer meeting coming up. Can you take a look at my calendar? You have a meeting with Sablecrest Robotics in 12 minutes, and you're meeting with Alex Kim, their CTO.”
Problem Agitation
“Hey, could you now update the CRM and put the meeting of today as a brief and the next steps? Let me pull the latest context and update your CRM. Sablecrest launched warehouse automation this morning. Expansion is active, security review is the blocker.”
Call to Action
“So that was a quick preview of our new real-time audio models coming to the OpenAI API. Voice can truly become the primary interface now, and we can't wait to see what you build with these new models. Thanks for watching. Smooth and clear. It felt natural and demo friendly.”
Synthesizable in Remotion · key components of the realtime-audio-demo-showcase template.
Dual-channel mirrored audio visualizer where green bars represent user input above the axis and grey/blue bars represent synthesized output below.
Map audio frequency buffer arrays into Remotion SVG paths; render upper rects using interpolate on input gain and lower rects on output buffer.
Small pill-shaped UI element showing spinning status indicator and monospace tool name upon execution.
Use spring animation to scale from 0.8 to 1 with an icon spin rotation loop based on frame count.
Purpose. Establish API release and present core capabilities.
Execution. Presenter speaking to camera from home/studio desk.
Purpose. Prove latency and multilingual natural conversation.
Execution. Interactive split waveform web application reacting to French and German speech.
Purpose. Demonstrate tool calling, preambles, and passive listening.
Execution. CRM software updating synchronously with smartphone voice interaction.
Purpose. Encourage developer adoption.
Execution. Presenter summary closing into OpenAI vector logomark.