Crate blazen_audio_tts 

Multi-backend text-to-speech surface for Blazen.

This crate provides one capability trait — TtsBackend — that extends blazen_audio::AudioBackend with synthesis and voice- management methods, plus a small set of concrete backends under backends:

• [backends::anytts::AnyTtsBackend] — fully local engine via any-tts (Kokoro-82M, VibeVoice, Qwen3-TTS). Gated by the anytts feature flag.
• backends::openai::OpenAiTtsBackend — HTTP client for OpenAI-compatible TTS servers (/v1/audio/speech + the de-facto /v1/voices/* extension surface). Gated by the openai feature flag (enabled by default).
• backends::piper::PiperBackend — reserved slot for a future Piper-ONNX backend; every method currently returns TtsError::Unsupported pending Wave 22 of the PR-AUDIO restructure.

Two provider wrappers compose backends with the same surface:

• TtsBackendHandle<B> — monomorphized over a concrete backend, used when the backend choice is fixed at compile time.
• DynTtsProvider — type-erased Arc<dyn TtsBackend>, used by the manager / pipeline layer.

Feature flags

Feature	Default	Description
openai	yes	Builds backends::openai::OpenAiTtsBackend.
anytts	no	Builds [backends::anytts::AnyTtsBackend].
bark	no	Builds [backends::bark::BarkBackend] (Suno-AI
		Bark: 3-stage AR transformer + EnCodec).
f5-tts	no	Builds [backends::f5::F5Backend] (SWivid F5-TTS:
		flow-matching DiT + Vocos vocoder).
engine	no	Deprecated alias for anytts; will be
		removed one release after the multi-backend
		restructure ships.

Re-exports

pub use backends::DEFAULT_MODEL;

pub use backends::DEFAULT_RESPONSE_FORMAT;

pub use backends::DEFAULT_VOICE;

pub use backends::OpenAiTtsBackend;

pub use backends::OpenAiTtsConfig;

pub use backends::OpenAiTtsSpeechRequest;

pub use backends::OpenAiTtsSpeechResponse;

pub use backends::PiperBackend;

Modules

backends

Concrete TtsBackend implementations.

Structs

DynTtsProvider

Erased TTS provider — used by the manager / pipeline layer when the backend is chosen at runtime.

TtsBackendHandle

Typed, monomorphized TTS provider holding a concrete backend B.

TtsOptions

Per-call options for TtsBackend::synthesize.

Enums

TtsError

Error type for any TTS backend in blazen-audio-tts.

TtsModel

Which underlying local TTS model to load when using the anytts backend.

Traits

TtsBackend

Capability trait for any TTS engine in the Blazen ecosystem.