tts

Piper TTS service implementation.

class pipecat.services.piper.tts.PiperTTSSettings(model: str | None | _NotGiven = <factory>, extra: dict[str, Any]=<factory>, voice: str | None | _NotGiven = <factory>, language: Language | str | None | _NotGiven = <factory>)

Bases: TTSSettings

Settings for PiperTTSService.

class pipecat.services.piper.tts.PiperTTSService(*, voice_id: str | None = None, download_dir: Path | None = None, force_redownload: bool = False, use_cuda: bool = False, settings: PiperTTSSettings | None = None, **kwargs)

Bases: TTSService

Piper TTS service implementation.

Provides local text-to-speech synthesis using Piper voice models. Automatically downloads voice models if not already present and resamples audio output to match the configured sample rate.

Settings

alias of PiperTTSSettings

__init__(*, voice_id: str | None = None, download_dir: Path | None = None, force_redownload: bool = False, use_cuda: bool = False, settings: PiperTTSSettings | None = None, **kwargs)

Initialize the Piper TTS service.

Parameters:

• voice_id –

  Piper voice model identifier (e.g. en_US-ryan-high).

  Deprecated since version 0.0.105: Use settings=PiperTTSService.Settings(voice=...) instead.

• download_dir – Directory for storing voice model files. Defaults to the current working directory.

• force_redownload – Re-download the voice model even if it already exists.

• use_cuda – Use CUDA for GPU-accelerated inference.

• settings – Runtime-updatable settings. When provided alongside deprecated parameters, settings values take precedence.

• **kwargs – Additional arguments passed to the parent TTSService.

can_generate_metrics() → bool

Check if this service can generate processing metrics.

Returns:

True, as Piper service supports metrics generation.

async run_tts(text: str, context_id: str) → AsyncGenerator[Frame, None]

Generate speech from text using Piper.

Parameters:

• text – The text to convert to speech.

• context_id – Unique identifier for this TTS context.

Yields:

Frame – Audio frames containing the synthesized speech and status frames.

class pipecat.services.piper.tts.PiperHttpTTSSettings(model: str | None | _NotGiven = <factory>, extra: dict[str, Any]=<factory>, voice: str | None | _NotGiven = <factory>, language: Language | str | None | _NotGiven = <factory>)

Bases: TTSSettings

Settings for PiperHttpTTSService.

class pipecat.services.piper.tts.PiperHttpTTSService(*, base_url: str, aiohttp_session: ClientSession, voice_id: str | None = None, settings: PiperHttpTTSSettings | None = None, **kwargs)

Bases: TTSService

Piper HTTP TTS service implementation.

Provides integration with Piper’s HTTP TTS server for text-to-speech synthesis. Supports streaming audio generation with configurable sample rates and automatic WAV header removal.

Settings

alias of PiperHttpTTSSettings

__init__(*, base_url: str, aiohttp_session: ClientSession, voice_id: str | None = None, settings: PiperHttpTTSSettings | None = None, **kwargs)

Initialize the Piper TTS service.

Parameters:

• base_url – Base URL for the Piper TTS HTTP server.

• aiohttp_session – aiohttp ClientSession for making HTTP requests.

• voice_id –

  Piper voice model identifier (e.g. en_US-ryan-high).

  Deprecated since version 0.0.105: Use settings=PiperHttpTTSService.Settings(voice=...) instead.

• settings – Runtime-updatable settings. When provided alongside deprecated parameters, settings values take precedence.

• **kwargs – Additional arguments passed to the parent TTSService.

can_generate_metrics() → bool

Check if this service can generate processing metrics.

Returns:

True, as Piper service supports metrics generation.

async run_tts(text: str, context_id: str) → AsyncGenerator[Frame, None]

Generate speech from text using Piper’s HTTP API.

Parameters:

• text – The text to convert to speech.

• context_id – Unique identifier for this TTS context.

Yields:

Frame – Audio frames containing the synthesized speech and status frames.