peteonrails/voxtype

Health v2 · maintenance only

Adoption not part of the score

1252 stars · 95 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Voxtype is a local, offline voice-to-text dictation tool for Linux with push-to-talk hotkey support across Wayland compositors like Hyprland, Sway, GNOME, and KDE. It runs multiple transcription engines (Cohere Transcribe, Whisper, Parakeet, and others) faster than realtime on CPU or GPU, outputting text at the cursor with no cloud or telemetry.

Use cases

• dictate text on linux with a push-to-talk hotkey
• offline speech-to-text on wayland
• voice typing in hyprland or sway
• transcribe audio locally without cloud services
• convert speech to text at cursor position on linux
• run whisper speech recognition on cpu in realtime
• record meeting transcripts with speaker attribution

When to choose

• you want fully local, private dictation on Linux with no cloud dependency
• you use a Wayland compositor like Hyprland, Sway, Niri, GNOME, or KDE
• you need fast CPU-based transcription with multiple engine choices
• you want meeting transcription with export to Markdown, SRT, or VTT

When to avoid

• you need voice-to-text on macOS or Windows
• you require X11-only workflows without evdev fallback support
• you want a graphical-first application rather than a daemon with TUI configuration

Facets

application · maturity active

speech-recognition cli accessibility speech-processing accessibility developer-tools cli rust voice-to-text push-to-talk wayland whisper dictation offline hyprland sway gnome kde transcription local-first command-line linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="peteonrails/voxtype")