Overview
Generates audio from input text using text-to-speech models. Supports multiple voices and output formats including mp3, opus, aac, flac, wav, and pcm.
Method Signature
func (r *AudioSpeechService) New(
ctx context.Context,
body AudioSpeechNewParams,
opts ...option.RequestOption,
) (*http.Response, error)
Request Parameters
The text to generate audio for. Maximum length is 4096 characters.
One of the available TTS models:
openai/tts-1 - Standard quality, fasteropenai/tts-1-hd - High definition, higher qualityopenai/gpt-4o-mini-tts - Latest model with additional features
The voice to use for generating audio. Supported voices:
alloyashballadcoralechofableonyxnovasageshimmerverse
The format to return the audio in. Supported formats:
mp3 - MPEG Audio Layer 3opus - Opus audio codecaac - Advanced Audio Codingflac - Free Lossless Audio Codecwav - Waveform Audio File Formatpcm - Pulse-Code Modulation
The speed of the generated audio. Select a value from 0.25 to 4.0.
1.0 is the default/normal speed- Values < 1.0 slow down the speech
- Values > 1.0 speed up the speech
Control the voice with additional instructions. Does not work with tts-1 or tts-1-hd. Only supported with gpt-4o-mini-tts.
The format to stream the audio in. Supported formats:
sse - Server-Sent Events (not supported for tts-1 or tts-1-hd)audio - Raw audio streaming
Response
Returns an http.Response containing the audio data stream. The response body should be read and saved to a file or streamed directly to the user.
Code Examples
Basic Text-to-Speech
package main
import (
"context"
"io"
"log"
"os"
dedalus "github.com/dedalus-labs/dedalus-sdk-go"
"github.com/dedalus-labs/dedalus-sdk-go/option"
)
func main() {
client := dedalus.NewClient(
option.WithAPIKey("your-api-key"),
)
ctx := context.Background()
response, err := client.Audio.Speech.New(ctx, dedalus.AudioSpeechNewParams{
Input: dedalus.F("Hello! This is a text-to-speech example."),
Model: dedalus.F("openai/tts-1"),
Voice: dedalus.F(dedalus.AudioSpeechNewParamsVoiceAlloy),
})
if err != nil {
log.Fatal(err)
}
defer response.Body.Close()
// Save to file
file, err := os.Create("speech.mp3")
if err != nil {
log.Fatal(err)
}
defer file.Close()
_, err = io.Copy(file, response.Body)
if err != nil {
log.Fatal(err)
}
log.Println("Audio saved to speech.mp3")
}
response, err := client.Audio.Speech.New(ctx, dedalus.AudioSpeechNewParams{
Input: dedalus.F("This speech will be faster and in FLAC format."),
Model: dedalus.F("openai/tts-1-hd"),
Voice: dedalus.F(dedalus.AudioSpeechNewParamsVoiceNova),
ResponseFormat: dedalus.F(dedalus.AudioSpeechNewParamsResponseFormatFlac),
Speed: dedalus.F(1.25), // 25% faster
})
Different Voices
voices := []dedalus.AudioSpeechNewParamsVoice{
dedalus.AudioSpeechNewParamsVoiceAlloy,
dedalus.AudioSpeechNewParamsVoiceEcho,
dedalus.AudioSpeechNewParamsVoiceFable,
dedalus.AudioSpeechNewParamsVoiceOnyx,
dedalus.AudioSpeechNewParamsVoiceNova,
dedalus.AudioSpeechNewParamsVoiceShimmer,
}
text := "Hello, this is a voice sample."
for i, voice := range voices {
response, err := client.Audio.Speech.New(ctx, dedalus.AudioSpeechNewParams{
Input: dedalus.F(text),
Model: dedalus.F("openai/tts-1"),
Voice: dedalus.F(voice),
})
if err != nil {
log.Printf("Error with voice %s: %v", voice, err)
continue
}
defer response.Body.Close()
// Save each voice sample
filename := fmt.Sprintf("voice_%s.mp3", voice)
file, _ := os.Create(filename)
io.Copy(file, response.Body)
file.Close()
}
With Instructions (GPT-4o-mini-tts)
response, err := client.Audio.Speech.New(ctx, dedalus.AudioSpeechNewParams{
Input: dedalus.F("Welcome to our podcast about AI technology."),
Model: dedalus.F("openai/gpt-4o-mini-tts"),
Voice: dedalus.F(dedalus.AudioSpeechNewParamsVoiceSage),
Instructions: dedalus.F("Speak in an enthusiastic and professional podcast host tone."),
})
Streaming Audio
response, err := client.Audio.Speech.New(ctx, dedalus.AudioSpeechNewParams{
Input: dedalus.F("This is a streaming audio example."),
Model: dedalus.F("openai/gpt-4o-mini-tts"),
Voice: dedalus.F(dedalus.AudioSpeechNewParamsVoiceAlloy),
StreamFormat: dedalus.F(dedalus.AudioSpeechNewParamsStreamFormatAudio),
})
if err != nil {
log.Fatal(err)
}
defer response.Body.Close()
// Stream audio data in chunks
buffer := make([]byte, 4096)
for {
n, err := response.Body.Read(buffer)
if err != nil && err != io.EOF {
log.Fatal(err)
}
if n == 0 {
break
}
// Process or play audio chunk
// ...
}
Voice Characteristics
- alloy - Neutral, versatile voice
- echo - Warm, engaging voice
- fable - Expressive, storytelling voice
- onyx - Deep, authoritative voice
- nova - Energetic, youthful voice
- shimmer - Soft, pleasant voice
- ash, ballad, coral, sage, verse - Additional voice options with unique characteristics
Best Practices
- Text Length: Keep input text under 4096 characters. For longer content, split into chunks.
- Format Selection: Use MP3 for general use, FLAC for highest quality, Opus for smallest file size
- Speed Adjustment: Use speed between 0.75-1.5 for natural-sounding results
- Voice Selection: Test different voices to find the best match for your use case
- Error Handling: Always check for errors and handle response cleanup properly