Text to speech | OpenAI API

The Audio API provides a speech endpoint based on our GPT-4o mini TTS (text-to-speech) model. It comes with 11 built-in voices and can be used to:

Narrate a written blog post
Produce spoken audio in multiple languages
Give realtime audio output using streaming

Here’s an example of the alloy voice:

Our usage policies require you to provide a clear disclosure to end users that the TTS voice they are hearing is AI-generated and not a human voice.

Quickstart

The speech endpoint takes three key inputs:

The model you’re using
The text to be turned into audio
The voice that will speak the output

Here’s a simple request example:

Generate spoken audio from input text

Python

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16import fs from "fs";
import path from "path";
import OpenAI from "openai";

const openai = new OpenAI();
const speechFile = path.resolve("./speech.mp3");

const mp3 = await openai.audio.speech.create({
  model: "gpt-4o-mini-tts",
  voice: "coral",
  input: "Today is a wonderful day to build something people love!",
  instructions: "Speak in a cheerful and positive tone.",
});

const buffer = Buffer.from(await mp3.arrayBuffer());
await fs.promises.writeFile(speechFile, buffer);

1
2
3
4
5
6
7
8
9
10
11
12
13from pathlib import Path
from openai import OpenAI

client = OpenAI()
speech_file_path = Path(__file__).parent / "speech.mp3"

with client.audio.speech.with_streaming_response.create(
    model="gpt-4o-mini-tts",
    voice="coral",
    input="Today is a wonderful day to build something people love!",
    instructions="Speak in a cheerful and positive tone.",
) as response:
    response.stream_to_file(speech_file_path)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34package main

import (
	"context"
	"io"
	"os"

	"github.com/openai/openai-go/v3"
)

func main() {
	client := openai.NewClient()
	response, err := client.Audio.Speech.New(context.Background(), openai.AudioSpeechNewParams{
		Model:        openai.SpeechModelGPT4oMiniTTS,
		Voice:        openai.AudioSpeechNewParamsVoiceUnion{OfAudioSpeechNewsVoiceString2: openai.String("coral")},
		Input:        "Today is a wonderful day to build something people love!",
		Instructions: openai.String("Speak in a cheerful and positive tone."),
	})
	if err != nil {
		panic(err)
	}
	defer response.Body.Close()

	file, err := os.Create("speech.mp3")
	if err != nil {
		panic(err)
	}
	if _, err := io.Copy(file, response.Body); err != nil {
		panic(err)
	}
	if err := file.Close(); err != nil {
		panic(err)
	}
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.core.http.HttpResponse;
import com.openai.models.audio.speech.SpeechCreateParams;
import java.io.IOException;
import java.nio.file.Files;
import java.nio.file.Path;
import java.nio.file.StandardCopyOption;

try (HttpResponse audio =
    client
        .audio()
        .speech()
        .create(
            SpeechCreateParams.builder()
                .model("gpt-4o-mini-tts")
                .voice("coral")
                .input("Today is a wonderful day to build something people love!")
                .instructions("Speak in a cheerful and positive tone.")
                .build())) {
  Files.copy(audio.body(), Path.of("speech.mp3"), StandardCopyOption.REPLACE_EXISTING);
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17using OpenAI.Audio;
#pragma warning disable OPENAI001

string key = Environment.GetEnvironmentVariable("OPENAI_API_KEY")!;
string model = "gpt-4o-mini-tts";
AudioClient client = new(model, key);

BinaryData audio = await client.GenerateSpeechAsync(
    "Today is a wonderful day to build something people love!",
    GeneratedSpeechVoice.Coral,
    new SpeechGenerationOptions
    {
        Instructions = "Speak in a cheerful and positive tone.",
    }
);

await File.WriteAllBytesAsync("speech.mp3", audio.ToArray());

1
2
3
4
5
6
7
8
9
10require "openai"

client = OpenAI::Client.new
audio = client.audio.speech.create(
  model: "gpt-4o-mini-tts",
  voice: "coral",
  input: "Today is a wonderful day to build something people love!",
  instructions: "Speak in a cheerful and positive tone."
)
File.binwrite("speech.mp3", audio.read)

1
2
3
4
5
6
7
8
9
10curl https://api.openai.com/v1/audio/speech \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini-tts",
    "input": "Today is a wonderful day to build something people love!",
    "voice": "coral",
    "instructions": "Speak in a cheerful and positive tone."
  }' \
  --output speech.mp3

1
2
3
4
5
6openai audio:speech create \
  --model gpt-4o-mini-tts \
  --voice coral \
  --instructions "Speak in a cheerful and positive tone." \
  --input "Today is a wonderful day to build something people love!" \
  --output speech.mp3

By default, the endpoint outputs an MP3 of the spoken audio, but you can configure it to output any supported format.

Text-to-speech models

For intelligent realtime applications, use the gpt-4o-mini-tts model, our newest and most reliable text-to-speech model. You can prompt the model to control aspects of speech, including:

Accent
Emotional range
Intonation
Impressions
Speed of speech
Tone
Whispering

Our other text-to-speech models are tts-1 and tts-1-hd. The tts-1 model provides lower latency, but at a lower quality than the tts-1-hd model.

Voice options

The TTS endpoint provides 13 built‑in voices to control how speech is rendered from text. Hear and play with these voices in OpenAI.fm, our interactive demo for trying the latest text-to-speech model in the OpenAI API. Voices are currently optimized for English.

alloy
ash
ballad
coral
echo
fable
nova
onyx
sage
shimmer
verse
marin
cedar

For best quality, we recommend using marin or cedar.

Voice availability depends on the model. The tts-1 and tts-1-hd models support a smaller set: alloy, ash, coral, echo, fable, onyx, nova, sage, and shimmer.

If you’re using the Realtime API, note that the set of available voices is slightly different—see the realtime conversations guide for current realtime voices.

Streaming realtime audio

The Speech API provides support for realtime audio streaming using chunk transfer encoding. This means the audio can be played before the full file is generated and made accessible.

Stream spoken audio from input text directly to your speakers

Python

1
2
3
4
5
6
7
8
9
10
11
12
13
14import OpenAI from "openai";
import { playAudio } from "openai/helpers/audio";

const openai = new OpenAI();

const response = await openai.audio.speech.create({
  model: "gpt-4o-mini-tts",
  voice: "coral",
  input: "Today is a wonderful day to build something people love!",
  instructions: "Speak in a cheerful and positive tone.",
  response_format: "wav",
});

await playAudio(response);

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21import asyncio

from openai import AsyncOpenAI
from openai.helpers import LocalAudioPlayer

openai = AsyncOpenAI()


async def main() -> None:
    async with openai.audio.speech.with_streaming_response.create(
        model="gpt-4o-mini-tts",
        voice="coral",
        input="Today is a wonderful day to build something people love!",
        instructions="Speak in a cheerful and positive tone.",
        response_format="pcm",
    ) as response:
        await LocalAudioPlayer().play(response)


if __name__ == "__main__":
    asyncio.run(main())

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27package main

import (
	"context"
	"io"
	"os"

	"github.com/openai/openai-go/v3"
)

func main() {
	client := openai.NewClient()
	response, err := client.Audio.Speech.New(context.Background(), openai.AudioSpeechNewParams{
		Model:          openai.SpeechModelGPT4oMiniTTS,
		Voice:          openai.AudioSpeechNewParamsVoiceUnion{OfAudioSpeechNewsVoiceString2: openai.String("coral")},
		Input:          "Today is a wonderful day to build something people love!",
		Instructions:   openai.String("Speak in a cheerful and positive tone."),
		ResponseFormat: openai.AudioSpeechNewParamsResponseFormatWAV,
	})
	if err != nil {
		panic(err)
	}
	defer response.Body.Close()
	if _, err := io.Copy(os.Stdout, response.Body); err != nil {
		panic(err)
	}
}

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46import com.openai.client.OpenAIClient;
import com.openai.client.okhttp.OpenAIOkHttpClient;
import com.openai.core.http.HttpResponse;
import com.openai.models.audio.speech.SpeechCreateParams;
import java.io.IOException;
import java.io.OutputStream;
import java.nio.file.Files;
import java.nio.file.Path;
import javax.sound.sampled.AudioFormat;
import javax.sound.sampled.AudioSystem;
import javax.sound.sampled.LineUnavailableException;
import javax.sound.sampled.SourceDataLine;

try (HttpResponse audio =
    client
        .audio()
        .speech()
        .create(
            SpeechCreateParams.builder()
                .model("gpt-4o-mini-tts")
                .voice("coral")
                .input("Today is a wonderful day to build something people love!")
                .instructions("Speak in a cheerful and positive tone.")
                .responseFormat(SpeechCreateParams.ResponseFormat.PCM)
                .streamFormat(SpeechCreateParams.StreamFormat.AUDIO)
                .build())) {
  AudioFormat format = new AudioFormat(24_000, 16, 1, true, false);
  String outputPath = System.getenv("OPENAI_EXAMPLE_AUDIO_OUTPUT_PATH");
  if (outputPath == null || outputPath.isBlank()) {
    try (SourceDataLine speakers = AudioSystem.getSourceDataLine(format)) {
      speakers.open(format);
      speakers.start();
      byte[] chunk = new byte[1024];
      int bytesRead;
      while ((bytesRead = audio.body().read(chunk)) != -1) {
        speakers.write(chunk, 0, bytesRead);
      }
      speakers.drain();
    }
  } else {
    try (OutputStream output = Files.newOutputStream(Path.of(outputPath))) {
      long bytes = audio.body().transferTo(output);
      System.out.println(bytes + " audio bytes");
    }
  }
}

1
2
3
4
5
6
7
8
9
10
11
12
13require "openai"

client = OpenAI::Client.new
audio = client.audio.speech.create(
  model: "gpt-4o-mini-tts",
  voice: "alloy",
  input: "Welcome to the OpenAI API.",
  response_format: :pcm,
  stream_format: :audio
)
while (chunk = audio.read(1_024))
  puts(chunk.bytesize)
end

1
2
3
4
5
6
7
8
9
10curl https://api.openai.com/v1/audio/speech \
  -H "Authorization: Bearer $OPENAI_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-4o-mini-tts",
    "input": "Today is a wonderful day to build something people love!",
    "voice": "coral",
    "instructions": "Speak in a cheerful and positive tone.",
    "response_format": "wav"
  }' | ffplay -i -

For the fastest response times, we recommend using wav or pcm as the response format.

Supported output formats

The default response format is mp3, but other formats like opus and wav are available.

MP3: The default response format for general use cases.
Opus: For internet streaming and communication, low latency.
AAC: For digital audio compression, preferred by YouTube, Android, iOS.
FLAC: For lossless audio compression, favored by audio enthusiasts for archiving.
WAV: Uncompressed WAV audio, suitable for low-latency applications to avoid decoding overhead.
PCM: Similar to WAV but contains the raw samples in 24kHz (16-bit signed, low-endian), without the header.

Supported languages

The TTS model generally follows the Whisper model in terms of language support. Whisper supports the following languages and performs well, despite voices being optimized for English:

Afrikaans, Arabic, Armenian, Azerbaijani, Belarusian, Bosnian, Bulgarian, Catalan, Chinese, Croatian, Czech, Danish, Dutch, English, Estonian, Finnish, French, Galician, German, Greek, Hebrew, Hindi, Hungarian, Icelandic, Indonesian, Italian, Japanese, Kannada, Kazakh, Korean, Latvian, Lithuanian, Macedonian, Malay, Marathi, Maori, Nepali, Norwegian, Persian, Polish, Portuguese, Romanian, Russian, Serbian, Slovak, Slovenian, Spanish, Swahili, Swedish, Tagalog, Tamil, Thai, Turkish, Ukrainian, Urdu, Vietnamese, and Welsh.

You can generate spoken audio in these languages by providing input text in the language of your choice.

Custom voices

Create an approved custom voice from a speaker’s consent recording and matching audio sample. See Custom voices for eligibility, recording requirements, consent phrases, and API requests.

Creating a voice

Follow Create a custom voice.

Using a voice during speech generation

Pass the created voice ID when generating speech. See the speech generation examples.

Realtime and audio overview

Choose the right path for voice agents, translation, transcription, and speech generation.

Audio and speech concepts

Review audio modalities, speech tasks, streaming, and request-based APIs.

Suggested

Get started

Core concepts

SDKs and CLI

Resources

Legacy APIs

Choose a model

Text and code

Prompting

Reasoning

Images and video

Realtime and audio

Specialized models

Agents API

Agents SDK

ChatKit

Search and retrieval

Connect tools and data

Build tool workflows

Computer and code

Media

GPT-Live

Realtime API

Build with voice

Connections

Audio processing

Go live

Performance and quality

Cost and throughput

Safety and governance

Infrastructure and access

Operations

Core concepts

Plan

Build

Test and publish

Conversion specs

Guides

Resources

Get started

Guides

File Upload

API

Measurement

Advertiser API

API Reference

Get started

Foundations

Explore

Available on

Releases

Workflows

Capabilities

Reference

Customization

Config file

Agent configuration

Extend ChatGPT and Codex

Linux

Windows

Development workflows

Extend and automate

Environments

Build with Codex

Third-party integrations

Reference

Permissions

Codex Security

Cyber safety

Getting started

ChatGPT Work

Identity and authentication

Workspace access, policy, and models

Plugin and connector controls

Usage, governance, and compliance

Deployment and model providers

Community

Blog

Community

Blog