Show HN: Wayland Speech-to-Text Tool
10 months ago
- #Wayland
- #speech-to-text
- #OpenAI
- Press a keybind to speak and get instant text output using OpenAI Whisper.
- Signal-driven operation with no GUI needed, outputs transcribed text to stdout.
- Works on Wayland desktops like Hyprland, Niri, GNOME, and KDE.
- Requires OpenAI API key for Whisper transcription and system packages like pipewire.
- Optional ydotool for direct typing keybindings.
- Install via AUR helper or GitHub Releases.
- Setup configuration with API key in ~/.config/waystt/.env.
- Test the application by running waystt and piping output.
- Use signals to trigger transcription or start waystt with options.
- Keybindings for direct typing or clipboard copy in Hyprland and Niri.
- Supports OpenAI Whisper (default) and Google Speech-to-Text providers.
- Configuration options for audio feedback, beep volume, and debug logging.
- Troubleshooting tips for audio recording and provider-specific issues.
- Development setup with cargo and git.
- Licensed under GPL v3.0.