sherpa-onnx-tts

To generate speech from text locally without cloud calls, use offline text-to-speech tools like sherpa-onnx. This approach synthesizes spoken audio directly on your machine, ensuring data privacy and eliminating network dependency for your TTS generation tasks.

To set up offline TTS with sherpa-onnx, configure two environment variables: SHERPA_ONNX_RUNTIME_DIR pointing to the installed runtime, and SHERPA_ONNX_MODEL_DIR pointing to your extracted voice model directory.

Yes, you can use local text-to-speech on macOS, Linux, and Windows. The sherpa-onnx runtime supports these operating systems, allowing you to download the correct runtime version and generate WAV audio files locally across different platforms.

To specify a different voice model for offline speech generation, use the SHERPA_ONNX_MODEL_FILE environment variable or the --model-file command line option. This allows you to select a specific .onnx model file for your audio output.

The best way to create voice prompts for demos offline is using a local CLI text-to-speech tool like sherpa-onnx. It processes text input into WAV audio files on your machine, providing consistent voice generation without relying on external cloud services.