earendil-works/pi-voice

transcription in pi

What it solves

Pi Voice provides a local, privacy-focused speech-to-text integration for the Pi terminal environment. It allows users to dictate text directly into their chat sessions and enables the AI agent to process local audio and video files without relying on external cloud transcription services.

How it works

It functions as an extension for the Pi platform, registering a global keyboard shortcut (defaulting to Ctrl+Alt+Z) to toggle recording. Once recorded, the audio is transcribed locally. It also provides a tool called transcribe_file that allows the AI agent to analyze audio or video files on your machine, utilizing FFmpeg for decoding.

Who it’s for

This tool is designed for users of the Pi terminal interface who want to interact with their AI agent via voice or need to transcribe local media files directly within their development workflow.

Highlights

  • Local-first transcription for privacy and offline capability.
  • Configurable keyboard shortcuts for hands-free interaction.
  • Integrated transcribe_file tool for agent-assisted media analysis.
  • Supports custom model selection and language settings via a dedicated TUI menu.

Related

  • Project
    NaomiProject/Naomi

    Naomi is an open‑source, always‑on voice‑assistant platform (Python, Linux/Raspberry Pi) that lets you control home devices, query information, and add custom “skills” via simple Python modules.

  • Project
    Elleo/pied

    Pied is a management tool for Linux that simplifies installing and configuring Piper neural text-to-speech voices for use with Speech Dispatcher.

  • Project
    kristoferlund/ostt

    OSTT is an open‑source, terminal‑native voice‑to‑text CLI for Linux/macOS. It records via a global hotkey, transcribes with local Whisper models or many cloud providers, and outputs to clipboard, files, stdout, or custom shell commands. Supports GPU acceleration, per‑model parameters, history/retry, and scriptable post‑processing actions.

  • Project
    zeroweight-ai/ZeroTTS

    ZeroTTS is a fast, accurate Vietnamese text-to-speech system that can clone a voice from just seconds of audio, runs in real time on a laptop CPU, and streams the first audio in about 70 milliseconds.