Voice
Choose Arbor, Breeze, Cove, Ember, Juniper, Maple, Sol, Spruce, or Vale under Settings → Voice.
Step-by-step setup
Open Voice from ChatGPT on the web, iOS, or Android, select Live in Voice settings, allow microphone access, and start talking. The exact options you see can depend on plan, region, workspace, and app version.
Last verified:
Open Settings → Voice and select Live. If it is missing, update ChatGPT and review the rollout checks in our troubleshooting guide.
Tap the Voice icon in the message bar, allow microphone access, choose a voice if prompted, and begin speaking.
Go to ChatGPT.com, select the Voice icon in the prompt window, grant browser microphone permission, and begin speaking.
Choose Arbor, Breeze, Cove, Ember, Juniper, Maple, Sol, Spruce, or Vale under Settings → Voice.
Set the language you speak most often to improve recognition, or ask ChatGPT to switch languages in the conversation.
Choose Instant, Medium, or High when the control is available; higher levels may respond more slowly.
While Live is active, type in the same chat or use the add button to attach a supported image. ChatGPT can continue answering in Voice. Available image types and limits depend on your plan and account.
Live cannot currently fetch files from the ChatGPT Library, though manual attachments may still be available. Video and screen sharing require Advanced Voice on eligible mobile accounts.
Enable Settings → Voice → Background conversations to keep talking while using another app or while the phone is locked.
On supported app versions, enable Start with Voice so a new or empty chat opens directly in Voice.
On supported iPhones, start Voice, continue a recent or pinned chat, or open a project from the CarPlay screen.
OpenAI says audio clips from Live and Advanced Voice are stored with the chat transcript and retained for 30 days. Deleting a chat also schedules its associated clips for deletion within 30 days, subject to documented safety, legal, and previously shared-data exceptions.
Audio or video clips are not used for model training unless you choose to share them through Data Controls. Transcripts and other files may be used when Improve the model for everyone is enabled, depending on your plan and settings.
GPT-Live-powered ChatGPT Voice now works in Codex in the macOS and Windows desktop apps. Start from a new empty Codex task and select Start new voice chat before sending the first message.
Open the Codex voice mode guideA Live conversation can end when you stop it, force-close the app, reach a usage limit, hit the two-hour per-conversation cap, or reach a long conversation's context limit.
See plan and session limitsYes, when Background conversations is enabled. It still ends at usage, session-length, force-close, or other documented stopping conditions.
Yes, but changing the selected voice starts a new voice call inside the same chat.
No at launch. Switch to Advanced Voice on an eligible iOS or Android account for supported video or screen sharing.