Package org.springframework.ai.openai

Class OpenAiAudioSpeechModel

java.lang.Object

org.springframework.ai.openai.OpenAiAudioSpeechModel

All Implemented Interfaces:

Model<SpeechPrompt,SpeechResponse>, StreamingModel<SpeechPrompt,SpeechResponse>, SpeechModel, StreamingSpeechModel

public class OpenAiAudioSpeechModel
extends Object
implements SpeechModel, StreamingSpeechModel

OpenAI audio speech client implementation for backed by OpenAiAudioApi.

Since:

1.0.0-M1

Author:

Ahmed Yousri, Hyunjoon Choi, Thomas Vitale, Jonghoon Park

See Also:

• OpenAiAudioApi

Constructor Summary

Constructors

Constructor	Description
OpenAiAudioSpeechModel(OpenAiAudioApi audioApi)	Initializes a new instance of the OpenAiAudioSpeechModel class with the provided OpenAiAudioApi.
OpenAiAudioSpeechModel(OpenAiAudioApi audioApi, OpenAiAudioSpeechOptions options)	Initializes a new instance of the OpenAiAudioSpeechModel class with the provided OpenAiAudioApi and options.
OpenAiAudioSpeechModel(OpenAiAudioApi audioApi, OpenAiAudioSpeechOptions options, org.springframework.retry.support.RetryTemplate retryTemplate)	Initializes a new instance of the OpenAiAudioSpeechModel class with the provided OpenAiAudioApi and options.

Method Summary

Modifier and Type	Method	Description
byte[]	call(String text)	Generates spoken audio from the provided text message.
SpeechResponse	call(SpeechPrompt speechPrompt)	Sends a speech request to the OpenAI TTS API and returns the resulting speech response.
reactor.core.publisher.Flux<SpeechResponse>	stream(SpeechPrompt speechPrompt)	Streams the audio response for the given speech prompt.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Methods inherited from interface org.springframework.ai.openai.audio.speech.StreamingSpeechModel

stream

Constructor Details

OpenAiAudioSpeechModel

public OpenAiAudioSpeechModel(OpenAiAudioApi audioApi)

Initializes a new instance of the OpenAiAudioSpeechModel class with the provided OpenAiAudioApi. It uses the model tts-1, response format mp3, voice alloy, and the default speed of 1.0.

Parameters:

audioApi - The OpenAiAudioApi to use for speech synthesis.

OpenAiAudioSpeechModel

public OpenAiAudioSpeechModel(OpenAiAudioApi audioApi,
 OpenAiAudioSpeechOptions options)

Initializes a new instance of the OpenAiAudioSpeechModel class with the provided OpenAiAudioApi and options.

Parameters:

audioApi - The OpenAiAudioApi to use for speech synthesis.

options - The OpenAiAudioSpeechOptions containing the speech synthesis options.

OpenAiAudioSpeechModel

public OpenAiAudioSpeechModel(OpenAiAudioApi audioApi,
 OpenAiAudioSpeechOptions options,
 org.springframework.retry.support.RetryTemplate retryTemplate)

Initializes a new instance of the OpenAiAudioSpeechModel class with the provided OpenAiAudioApi and options.

Parameters:

audioApi - The OpenAiAudioApi to use for speech synthesis.

options - The OpenAiAudioSpeechOptions containing the speech synthesis options.

retryTemplate - The retry template.

Method Details

call

public byte[] call(String text)

Description copied from interface: SpeechModel

Generates spoken audio from the provided text message.

Specified by:

call in interface SpeechModel

Parameters:

text - the text message to be converted to audio

Returns:

the resulting audio bytes

call

public SpeechResponse call(SpeechPrompt speechPrompt)

Description copied from interface: SpeechModel

Sends a speech request to the OpenAI TTS API and returns the resulting speech response.

Specified by:

call in interface Model<SpeechPrompt,SpeechResponse>

Specified by:

call in interface SpeechModel

Parameters:

speechPrompt - the speech prompt containing the input text and other parameters

Returns:

the speech response containing the generated audio

stream

public reactor.core.publisher.Flux<SpeechResponse> stream(SpeechPrompt speechPrompt)

Streams the audio response for the given speech prompt.

Specified by:

stream in interface StreamingModel<SpeechPrompt,SpeechResponse>

Specified by:

stream in interface StreamingSpeechModel

Parameters:

speechPrompt - The speech prompt containing the text and options for speech synthesis.

Returns:

A Flux of SpeechResponse objects containing the streamed audio and metadata.