WOLFRAM

Wolfram Language & System Documentation Center

"ElevenLabs" (Service Connection)

This service connection requires LLM access »
Use the ElevenLabs API with the Wolfram Language.

Connecting & Authenticating

ServiceConnect["ElevenLabs"] creates a connection to the ElevenLabs API. If a previously saved connection can be found, it will be used; otherwise, a new authentication request will be launched.
Use of this connection requires internet access and an ElevenLabs account.

Requests

ServiceExecute["ElevenLabs","request",params] sends a request to the ElevenLabs API, using parameters params. The following give possible requests.

Synthesize Speech

Request:

"ListVoices" — list of available voices for speech synthesis

Request:

"TextToSpeech" — synthesize speech from text

Parameters:
  • "Text"(required)the text to synthesize
    "Voice"Automaticthe name or specification for the voice to be used
    "VoiceSettings"Automaticassociation of settings for the voice to be used
    "ModelID"Automaticname of the model to be used
    "OptimizeStreamingLatency"Automaticinteger between 0 and 4 to control latency

    • Examples

    open all close all

    Basic Examples  (1)

    Create a new connection:

    Wolfram Language code: el = ServiceConnect["ElevenLabs"]

    List all available voices:

    Wolfram Language code: ServiceExecute[el, "ListVoices"]//Dataset

    Synthesize a piece of text:

    Wolfram Language code: ServiceExecute[el, "TextToSpeech", {"Text" -> "hello there!"}]

    Scope  (3)

    Choose a voice at random:

    Wolfram Language code: voice = RandomChoice[ServiceExecute["ElevenLabs", "ListVoices"]]

    Synthesize a piece of text with that voice:

    Wolfram Language code: ServiceExecute["ElevenLabs", "TextToSpeech", {"Text" -> "hello there!", "Voice" -> voice}]

    Specify the voice just by its name:

    Wolfram Language code: ServiceExecute["ElevenLabs", "TextToSpeech", {"Text" -> "hello there!", "Voice" -> voice["Name"]}]

    Each result will be slightly different since it is produced using a neural network.

    Specify settings to be used for the synthesis:

    Wolfram Language code: ServiceExecute["ElevenLabs", "TextToSpeech", {"Text" -> "hello there!", "VoiceSettings" -> <|"similarity_boost" -> 1, "stability" -> 0, "style" -> 1|>}]

    Increasing the "OptimizeStreamingLatency" parameter will return faster results at the cost of quality:

    Wolfram Language code: ServiceExecute["ElevenLabs", "TextToSpeech", {"Text" -> "hello there! This will be a longer sentence", "OptimizeStreamingLatency" -> 4}]//AbsoluteTiming
    Wolfram Language code: ServiceExecute["ElevenLabs", "TextToSpeech", {"Text" -> "hello there! This will be a longer sentence", "OptimizeStreamingLatency" -> 1}]//AbsoluteTiming