Moonshine Voice (Go) brings low-latency, on-device speech recognition to the Go ecosystem. Providing idiomatic bindings and high-level audio abstractions around the Moonshine Voice C library, it enables developers to build local voice assistants and streaming transcription pipelines with zero cloud latency.

Key Features

  • Idiomatic Go Interface: Wraps the native C Moonshine Voice engine with clean Go channels, contexts, and reader/writer streaming interfaces.
  • Fast Local Transcription: Runs compact ONNX acoustic models locally on CPU or Apple Silicon neural accelerators.
  • Microphone & File Ingestion: Built-in CLI utilities for real-time microphone capture, chunked streaming, and WAV audio transcription.
  • Voice Agent Scaffolding: Designed for seamless integration into multi-modal agent harnesses and local voice loops.

Explore the repository on GitHub: github.com/ghchinoy/moonshine-go.