How to Use the ChatGPT Text-to-Speech (TTS) Model

Want to hear ChatGPT talk? OpenAI’s text-to-speech (TTS) model turns written responses into expressive, human-like speech. Whether you’re creating bedtime stories, voiceovers, or just experimenting with audio, it opens up new ways to interact with AI.

What is the ChatGPT text-to-speech (TTS) model?

ChatGPT’s text-to-speech (TTS) feature converts any text into natural-sounding audio directly within the ChatGPT interface. Here’s how to use it:

Quick start guide:

1. Mobile App: Open ChatGPT mobile app → Settings → New Features → Enable Voice Conversations → Tap headphone icon
2. Web Interface: Look for the speaker icon below any ChatGPT response and click to play audio
3. Voice Selection: Choose from 5 built-in voices (Breeze, Sol, Ember, Arbor, Vale, Spruce)

Requirements:

• ChatGPT Plus subscription for mobile voice features
• Free users can use the web speaker button for basic text-to-speech

The ChatGPT TTS model is designed to transform content into expressive, lifelike speech with multiple voices and real-time audio playback with no coding required.

Here’s what makes it stand out:

• Multiple voices and speaking styles: Choose from voices like Ember (warm), Spruce (confident), and Sol (bright)

• Real-time audio playback: Instantly generates audio you can play back

• No coding required: Direct integration within ChatGPT interface

In this tutorial, we’ll build a voice-based storytelling project using ChatGPT text-to-speech. Let’s get started.

Build a voice-based storyteller using ChatGPT’s TTS model

This storyteller will read out the content given, and you’ll be able to control what it says, how it says it, and the voice it uses. Here’s what we will do:

1. Choose a theme
2. Write a short script
3. Pick a voice
4. Generate audio

Step 1: Pick a theme

Before generating audio, we need a theme that will act as the purpose of this storyteller. It’ll decide what the listener hears, how it sounds, and why it matters. Ask yourself, “What kind of voice experience do I want to create? Who is it for?”

Here are some example themes that you can adapt:

• Bedtime stories: Short tales with a warm, soothing voice for kids to calm down.
• Daily motivation: Uplifting quotes or affirmations delivered with clarity and energy.
• Did-you-know facts: Bite-sized trivia with a friendly or quirky tone.
• Mindfulness moments: Simple breathing prompts or calming reminders in a gentle voice.
• Quick news bites: Snappy updates or summaries in a neutral, steady tone.

Choose one that meets your goals.

For this walkthrough, let’s go with bedtime stories. With the theme decided, it’s time to bring it to life with a script.

Step 2: Write a sample script

A script is the piece of text to be spoken out loud. It could be a calming intro, a gentle story, or even a soothing poem for bedtime stories. Keep it clear and age-appropriate.

Here are some tips if you are writing the entire story:

• Use natural language and write like you talk.
• Contractions, pauses, and simple phrasing sound better.
• Match the tone to your theme.

But if you want a shortcut, you can either start ChatGPT and let it generate the story for you or ask ChatGPT to write the entire story. Just describe your theme and audience clearly. For example, we can have the following prompt for our theme:

Write a gentle and soothing bedtime story for a 6-year-old. The story should use calm, simple language and be easy to follow. It should feature a kind animal character and end with the character peacefully sleeping. Keep the tone warm and comforting. 

Copy to clipboard
Copy to clipboard

That’ll give you a solid draft. You can then tweak it to match a specific idea. Here’s what a sample output looks like:

🌙 "Luna the Little Fox" 🦊 

Once upon a time, in a quiet forest where the trees whispered softly in the breeze, there lived a gentle little fox named Luna. 

Luna had the softest fur, the color of autumn leaves, and big sleepy eyes that sparkled like stars. She loved to explore during the day—chasing butterflies, listening to the songs of the birds, and watching the clouds float by like fluffy pillows in the sky. 

But tonight, the sky was turning deep blue, and tiny stars were peeking out one by one. 

Luna yawned. She felt a little sleepy, but she wanted to say goodnight to all her forest friends first. 

She padded gently to the pond, where the frogs were humming their evening tune. 

"Goodnight, Froggies," Luna whispered with a smile. 

"Goodnight, Luna," they croaked back, their voices low and slow. 

She tiptoed past the tall grass, where the crickets chirped their lullaby. 

"Goodnight, Crickets," Luna said kindly. 

"Goodnight, Luna," the crickets replied. 

Next, she visited an old oak tree where her friend Ollie the Owl lived. 

"Goodnight, Ollie," said Luna softly. 

Ollie blinked his big eyes and hooted, "Sleep well, little fox. Sweet dreams await."

Feeling warm and cozy inside, Luna walked back to her den. The moon was shining gently above, wrapping the forest in silver light. 

She curled up in her soft bed of moss and leaves, her tail tucked around her like a blanket. 

A breeze blew through the trees, singing a lullaby just for her. 

"Shhh," it seemed to say. "Close your eyes, little fox. The night is here to keep you safe." 

Luna yawned once more, her eyes growing heavy. 

She thought of the butterflies and the birds and the stars in the sky. 

And before she knew it... 

she was fast asleep, dreaming happy dreams under the twinkling stars. 

The forest was quiet. All was well. And Luna the little fox slept soundly through the night.🌟 
Goodnight. 

Copy to clipboard
Copy to clipboard

Once the script is ready, we need to decide how it sounds.

Step 3: Choose a voice

Voice is everything. It sets the mood and brings your words to life. The same line can feel cozy, serious, or exciting depending on the voice behind it.

When using ChatGPT text-to-speech, voice selection significantly impacts your audio output. The ChatGPT TTS feature offers multiple voice personalities to match your content’s tone and audience.

To play around with different voices, go to your Account > Settings > Voice and preview the different voices. Here are a few voice styles you can try:

• Ember: Warm and expressive. Great for reflective thoughts or gentle motivation.
• Spruce: Confident and calm. Ideal for narration, general updates, or storytelling.
• Sol: Bright and friendly. Perfect for fun facts, light prompts, or casual stories.
• Arbur: Deep and relaxed. Works for bedtime stories, meditations, or advice.
• Vale: Smooth and clear. Good for anything professional, polished, or instructional.
• Breeze: Light and cheerful. Great for kids’ content or daily reminders.

You can test each voice to see what feels right for your theme.

Step 4: Play or download your voice experience

Now, to convert this content using ChatGPT’s text-to-speech, you can either:

• Play it directly: There’s a play button below this story that you can select and let ChatGPT read it out loud using a chosen voice.

• Download it as an MP3: Generate the audio in a downloadable format and then save it for later use.

Here’s a prompt that we can use to generate a downloadable audio:

Create an MP3 file of this bedtime story. 

Copy to clipboard
Copy to clipboard

Here’s a sample output generated:

And there it is, your own custom voice story, written by you, voiced by AI, ready to play or share!

Conclusion

We just walked through the steps of crafting a meaningful voice experience from choosing a theme, writing a simple script, to generating an audio file you can play or download anytime. This project showed how you can use ChatGPT text-to-speech (TTS) to bring stories and ideas to life with a voice style that feels personal and engaging.

Want to build more with AI? Explore Codecademy’s Learn Prompt Engineering course to sharpen your prompting skills and create more interactive AI projects.

Frequently asked questions

1. What are the models of OpenAI text-to-speech?

OpenAI’s text-to-speech (TTS) system uses six voices: Ember, Breeze, Sol, Echo, Fable, and Onyx, built on a model called Whisper for speech recognition and TTS for voice generation.

2. How to get ChatGPT speech?

You can use ChatGPT’s built-in voice feature (in the mobile app), or convert text to speech using a prompt in the desktop version, then play or download the audio.

3. Is ChatGPT voice free?

The voice feature is free for ChatGPT users on mobile (iOS and Android). Voice generation through text prompts and playback is also available via GPT-4o for Plus users.

4. How to download ChatGPT voice?

Use a prompt like “Convert this story to an mp3 audio file” and ChatGPT can generate downloadable audio if the feature is enabled in your plan.

5. Does ChatGPT 4 have text-to-speech?

Yes. GPT-4o, the latest version, includes real-time voice capabilities and supports text-to-speech output in various voices.