What is ChatGPT Sky Voice and how it works
ChatGPT Sky Voice is a real-time voice mode that lets you speak with GPT‑4o via your microphone and hear GPT‑4o’s replies through your speakers or headphones. It’s designed for natural turn‑taking, so you can interrupt and ask follow‑ups while GPT‑o responds conversationally. Sky Voice uses automatic speech recognition (ASR) to transcribe your words, a language model to understand intent and context, and text‑to‑speech (TTS) to produce fluent, expressive responses. The mode is available in ChatGPT apps and web interface when you pick the GPT‑4o model and switch to Voice, provided your account has access and your device supports microphone and speaker input.
Requirements and how to enable Sky Voice in ChatGPT
To use Sky Voice, you need:
- A ChatGPT account with access to GPT‑4o (Plus, Team, or Enterprise depending on availability).
- The latest version of the ChatGPT app or a modern browser with a supported microphone and speakers or headphones.
- Permission for ChatGPT to access your microphone and, if desired, camera (camera is optional and not used for voice processing).
To enable it:
- Open ChatGPT and select GPT‑4o (or “GPT‑4o mini” where offered) in the model picker.
- Start a message or conversation, then choose the Voice tab (usually a loudspeaker icon).
- Allow microphone and speaker permissions when prompted, then tap the circular microphone button to speak.
- Tap the button again to stop speaking and wait for GPT‑4o’s audio reply.