Refactor speech processing: add type hint for `_process_active_voice_frame` and replace `max()` with `np.max()` for consistency.

Refactor TTS audio generation: rename queues for clarity, update `TTSAudioGenerator` initialization, and enhance docstrings for better maintainability.

Refactor TTS architecture: implement runtime interface, TTS manager, universal registry, and factory pattern to support multiple engines.

Refactor voice model structure: extract MoYoYo-specific configurations and introduce universal TTS registry.

Remove unused `prompt_semantic` and `reference_spec` configuration parameters from voice model definitions.

Add thread readiness checks and is_ready property across services

Using FunASR quantized model.

Add multilingual support and optimize LLM pipeline configuration.

Add descriptions for Chinese voice models.

Remove unused configuration parameters and conversation templates.

Integrate FunASR service.

First commit.