What problem does it solve? Cloud text-to-speech services require internet access, send your text to third-party servers, and incur per-request costs. This Skill converts text to speech entirely offline using the sherpa-onnx runtime, keeping all audio generation local and private. ## Core Features & Use Cases - Offline Text-to-Speech: Synthesizes WAV audio from text using the sherpa-onnx CLI with VITS/Piper voice models, with no network connection required. - Cross-Platform Runtime: Provides prebuilt runtime downloads for macOS, Linux, and Windows, plus a downloadable en_US Lessac voice model. - Configurable Voices: Supports swapping in any model from the sherpa-onnx tts-models release and overriding model, tokens, and data files via environment variables or CLI flags. - Use Case: A privacy-focused voice assistant running in an isolated environment can speak responses aloud by generating tts.wav files locally without sending any text to a cloud API. ## Quick Start Ask the agent to run the sherpa-onnx-tts wrapper with your text, for example: generate a WAV file saying "Hello from local TTS" using the sherpa-onnx-tts skill.
