How to change ChatGPT's voice
ChatGPT offers voice conversation on mobile apps and the web, and you can switch between five different voices in the settings. The steps differ slightly depending on whether you're using a phone or a computer, but the voice options themselves are the same across all platforms.
On iPhone or Android, open the ChatGPT app, tap your profile icon in the bottom right, select Settings, then tap Voice. You'll see five voice options — Breeze, Cinnamon, Ember, Juniper, and Sage — each with a play button so you can hear a sample before choosing. Tap the one you want, and it becomes your default for all voice conversations going forward.
On the web version at openai.com, start a voice conversation by clicking the headphone icon in the message box. Once the conversation is active, click the three dots in the top right of the chat window and select Voice Settings. The same five voices appear, and you can preview each one before selecting your preference.
Key Takeaways
- ChatGPT offers five distinct voices — Breeze, Cinnamon, Ember, Juniper, and Sage — available on both mobile apps and the web.
- On mobile, voice settings are found in your profile menu under Settings, then Voice.
- On the web, you access voice settings by starting a voice conversation and clicking the three-dot menu in the chat window.
- Your chosen voice becomes the default for all future voice conversations until you change it again.
- You can preview each voice before selecting it to hear how it sounds.
Where to find voice settings on each platform
The location of voice settings depends on which device or browser you're using. Mobile users have the quickest path — the settings are grouped with other account preferences in one place. Web users need to start a conversation first, since voice settings only appear once a voice chat is active.
On iPhone, the ChatGPT app stores voice settings in the same menu as language preferences and notification controls. After tapping your profile icon and selecting Settings, you'll see Voice listed alongside other options. On Android, the layout is identical — profile icon, Settings, then Voice.
On the web, voice settings are conversation-specific rather than account-wide. This means you can technically use different voices in different conversations if you want, though most people set one and leave it. The three-dot menu appears in the top right corner of the chat window only when a voice conversation is active.
What each voice sounds like and how to preview them
Each of the five voices has a distinct tone and pace. Breeze is calm and measured, Cinnamon is warm and conversational, Ember is energetic and clear, Juniper is bright and friendly, and Sage is thoughtful and deliberate. The differences are noticeable enough that you'll likely have a preference within a few listens.
Every platform lets you hear a sample before committing. On mobile, tap the play button next to each voice name. On the web, the play button appears next to each option in the Voice Settings menu. Listen to at least two or three before deciding — the voice you'll hear during long conversations should feel natural to you, not distracting.
If you change your mind later, you can switch voices at any time using the same steps. There's no limit to how often you can change, so you can experiment until you find what works best for your listening style.
Voice requirements and device compatibility
Voice conversations require a ChatGPT Plus subscription or a free account with voice access enabled. Not all free accounts have voice turned on yet — OpenAI has been rolling it out gradually. If you don't see a voice option in your settings, your account may not have access yet, and you'll need to check back later or upgrade to Plus.
On mobile, the ChatGPT app for iPhone and Android both support all five voices equally. On the web, voice works in Chrome, Safari, Edge, and Firefox, though the experience is most stable in Chrome. If you're using an older browser version, updating it may be necessary to access voice features.
Voice conversations also require a working microphone and speaker or headphones. If you're on a device without a microphone — like a tablet without audio input — you won't be able to use voice, though you can still read ChatGPT's text responses.
Changing your voice mid-conversation
You cannot change the voice in the middle of an active conversation. If you start a chat with Breeze and decide you want Cinnamon instead, you'll need to end the current conversation and start a new one with your preferred voice selected.
This is a minor inconvenience, but it means you should pick a voice before beginning a longer conversation. If you're just testing voices out, start short conversations with each one to get a feel for how it sounds over time, rather than switching partway through.
Your voice choice persists across conversations once you've set it. If you change your default voice in settings, every new conversation will use the new voice unless you change it again.
Troubleshooting voice settings that won't change
If you select a voice and it doesn't seem to take effect, the most common cause is that you haven't fully saved the change. On mobile, make sure you tap the voice name itself, not just the play button. On the web, confirm that the voice you selected shows a checkmark or highlight next to it.
If the voice still doesn't change after you've confirmed the selection, try closing the app or browser tab completely and reopening it. This forces the app to reload your settings from the server. On mobile, you can also try signing out of your account and signing back in.
If you're on the web and voice settings don't appear at all, make sure you've actually started a voice conversation by clicking the headphone icon. Voice settings only show up once a conversation is active. If the headphone icon isn't visible, your account may not have voice access yet.
Using voice in different contexts
Voice conversations work best for quick questions, brainstorming, and situations where typing isn't convenient — like when you're cooking, driving, or working with your hands. The five voices are designed to be pleasant to listen to for these shorter interactions, though they work fine for longer conversations too.
If you find yourself having very long voice conversations regularly, experiment with which voice feels least fatiguing to listen to over time. Some people find faster voices like Ember better for information-heavy chats, while others prefer the steadier pace of Sage or Breeze for complex topics.
You can also use voice to have ChatGPT read back text responses. After you type a message, ChatGPT will respond with text, and you can tap or click to have it read aloud in your chosen voice. This is separate from voice conversations where you speak to ChatGPT — it's just text-to-speech playback.
Frequently Asked Questions
Can I use different voices in different conversations?
On the web, yes — voice settings are per-conversation, so you could theoretically use Breeze in one chat and Cinnamon in another. On mobile, your voice choice is account-wide, so all conversations use the same voice unless you change your default setting.
What if I don't like any of the five voices?
You're limited to the five voices ChatGPT currently offers. OpenAI may add more voices in the future, but for now, these are your only options. If none feel right, you can always use text-based chat instead and skip voice entirely.
Do I need ChatGPT Plus to use voice?
No, but voice access is rolling out gradually to free accounts. If you have a free account and don't see voice options, your account doesn't have access yet. Plus subscribers have may provide access to voice features.
Will my voice choice sync across my devices?
On mobile, your voice preference syncs across iPhone and Android if you're signed into the same account. On the web, voice settings are conversation-specific, so they don't carry over from one browser to another automatically.
Can I change the voice speed or pitch?
No, ChatGPT doesn't offer controls for speed or pitch. The five voices are fixed — you can only choose between them, not customize how they sound.