sherpa-onnx-tts

Offline text-to-speech synthesis can run locally using sherpa-onnx to generate speech on-device. This ensures privacy and low latency by synthesizing audio without network access, saving the output directly as a WAV file.

To set up a local TTS runtime, you need to configure environment variables for SHERPA_ONNX_RUNTIME_DIR and SHERPA_ONNX_MODEL_DIR. This directs the wrapper to the correct OS-specific runtime and model downloads for speech synthesis.

Yes, local speech synthesis supports generating voice prompts for offline assistants. It operates without cloud services, enabling on-device voice generation for offline demos and reading long text aloud in applications.

On-device speech synthesis works completely offline without an internet connection. By using sherpa-onnx locally, it processes text and generates speech audio entirely on your device, ensuring data privacy and low latency.

You can save synthesized speech output as a WAV file by using the sherpa-onnx-tts wrapper. The wrapper executes local speech synthesis and writes the resulting audio directly to disk for offline use.

Offline TTS is used instead of cloud-based speech synthesis to guarantee privacy and achieve low latency. It processes speech generation locally on-device, making it suitable for environments without network access.