Fact · Speech recognition
OpenAI's Realtime API voice activity detection, its default server-side turn detection for models that support it, uses periods of silence to split the audio into turns.
Checked against its source on October 17, 2026
In the source's own words
“It uses periods of silence to automatically chunk the audio.”
Voice activity detection (VAD), OpenAI Realtime API, OpenAI, OpenAI Developers, n.d., Section on voice activity detection (server_vad mode).
Open the original- What it measures
- The guide describes server voice activity detection as the default mode for speech-to-speech sessions and for transcription sessions on models that support turn detection.
- What it does not show
- Describes OpenAI's Realtime API. Other services use their own endpointing rules, and the guide does not describe how the detector works internally.
Source
Cited on
- Voice Assistants vs Dictation: Siri, Alexa, GeminiGuide
- AI Voice Agents: How They Work and Where They FailGuide
- How do you connect speech to text to an LLM?Question
- How does speech recognition know when you stop talking?Question
- How fast does a voice AI agent need to respond?Question
- Turn detectionGlossary
- Voice activity detection (VAD)Glossary