Moonshine Voice (Rust) provides safe, ergonomic, and high-performance Rust bindings for the Moonshine Voice speech-to-text engine. Featuring built-in multi-format audio decoding via Symphonia and static linking capabilities, it enables single-binary deployment of on-device voice transcription applications.

Key Features

  • Zero-Dependency Audio Decoding: Built-in support for streaming and decoding MP3, WAV, and FLAC audio files without external runtime libraries.
  • Thread-Safe Speech Pipeline: Leverages Rust's ownership model for concurrent audio pre-processing and model inference.
  • Cross-Platform Compilation: Supports macOS (including Metal acceleration), Linux, and embedded ARM targets.
  • Agent-Ready CLI: Includes a standalone binary for fast terminal transcription and pipe-friendly streaming output.

Explore the crate and examples on GitHub: github.com/ghchinoy/moonshine-rs.