MediaPipeTasksAudio Framework Reference

TextToSpeech

final class TextToSpeech : @unchecked Sendable

MediaPipe TextToSpeech task interface for Swift (iOS / macOS).

• init(options:)

  Creates and initializes a TextToSpeech instance configured with TextToSpeechOptions.

  Declaration

  Swift

  public init(options: TextToSpeechOptions) throws

• listVoices()

  Queries available voices for this initialized TextToSpeech instance.

  Declaration

  Swift

  public func listVoices() throws -> [TtsVoiceInfo]

• synthesize(_:options:)

  Synchronously synthesizes text and returns the TtsAudioResult.

  Declaration

  Swift

  public func synthesize(_ text: String, options: TtsSynthesisOptions? = nil) throws -> TtsAudioResult

• synthesizeAsync(_:options:)

  Asynchronously synthesizes text off the main thread with structured concurrency cancellation.

  Declaration

  Swift

  public func synthesizeAsync(_ text: String, options: TtsSynthesisOptions? = nil) async throws -> TtsAudioResult

• synthesizeStream(_:options:)

  Streams synthesized PCM TtsAudioChunk values as each sentence completes.

  Declaration

  Swift

  public func synthesizeStream(_ text: String, options: TtsSynthesisOptions? = nil) -> AsyncThrowingStream

• openStream(_:options:finishInput:)

  Opens a TtsStream for incremental LLM text streaming (pushText + finishInput) or pull iteration (readNext).

  Declaration

  Swift

  public func openStream(_ initialText: String = "", options: TtsSynthesisOptions? = nil, finishInput: Bool = false) throws -> TtsStream

• cancel()

  Thread-safe, lock-free barge-in cancellation of any in-flight synthesis.

  Declaration

  Swift

  public func cancel()