Voxtype vs hyprwhspr

Two Whisper-based tools for Hyprland users. Both offer Waybar integration and push-to-talk. Which fits your setup?

At a Glance

Aspect	Voxtype	hyprwhspr
Engine	Whisper (whisper.cpp)	Whisper (pywhispercpp)
Language	Rust	Python
Target Distro	Any Linux	Arch/Omarchy
Compositors	Any Wayland	Hyprland-focused
GPU Acceleration	Vulkan, CUDA, Metal, ROCm	CUDA, ROCm (6.x only)
Waybar Integration	Yes	Yes
Systemd Service	Yes	Yes
Audio Feedback	Yes (customizable themes)	Yes
Desktop Notifications	Yes (notify-send)	No (audio only)
Word Replacements	Yes	Yes (via WhisperTux)
Spoken Punctuation	Yes (opt-in)	Yes
Interactive Setup	Yes (`voxtype setup model`)	Yes (`hyprwhspr setup`)
Cloud Fallback	No (offline only)	Yes (OpenAI, Groq APIs)

What They Share

Both tools have excellent feature parity:

Whisper-powered - Both use OpenAI's Whisper for high-accuracy transcription
Waybar integration - Status indicators for tiling WM users
Systemd user service - Starts at login, always ready
Audio feedback - Know when recording starts/stops
Push-to-talk - Hold hotkey to record
ydotool integration - Text appears at cursor
Word replacements - Custom text substitutions for misheard words
Spoken punctuation - Say "period" to get ".", "open paren" for "(", etc.
Interactive setup - Guided model selection and download

Critical Differences

Distro & Compositor Support

Voxtype works on any Linux distribution and any Wayland compositor. It uses evdev for hotkey detection, which is compositor-agnostic.

hyprwhspr is designed specifically for Arch Linux and Omarchy. While it may work on other distros, it's optimized for the Hyprland ecosystem and may require extra setup elsewhere.

Architecture

Voxtype is a single Rust binary with no runtime dependencies beyond system libraries. Install once, run forever.

hyprwhspr is a Python application that relies on pywhispercpp and other Python dependencies. It offers more flexibility (cloud API fallbacks, word overrides) but has a larger dependency footprint.

GPU Acceleration

Voxtype supports Vulkan (works on AMD, NVIDIA, Intel), CUDA, Metal, and ROCm via compile-time features.

hyprwhspr supports CUDA and ROCm, but notes that ROCm 7.x has compatibility issues and falls back to CPU. If you're on AMD with recent ROCm, verify compatibility.

Feature Comparison

What hyprwhspr Does Better

Cloud API fallback - Can use OpenAI or Groq APIs when local transcription is too slow
Parakeet support - Alternative to Whisper using NVIDIA's Parakeet-v3 model (requires NeMo)

What Voxtype Does Better

Any compositor - Works on GNOME, KDE, Sway, Hyprland, River, and more
Any distro - Packages for Arch, Debian, Fedora, or build from source
Single binary - No Python environment to manage
Desktop notifications - Visual feedback via notify-send, not just audio
Privacy-first - No cloud options means your voice never leaves your machine
Broader GPU support - Vulkan works across GPU vendors (including ROCm 7.x)
large-v3-turbo model - Fast multilingual model (1.6 GB) with excellent accuracy

Setup Comparison

Voxtype on Arch

paru -S voxtype
voxtype setup model     # Interactive model selection
voxtype setup systemd   # Install systemd service
voxtype setup waybar    # Get Waybar config (optional)

hyprwhspr on Arch

yay -S hyprwhspr
hyprwhspr setup
# Follow interactive prompts

The Verdict

Choose Voxtype if you want a single binary that works on any Wayland compositor and any Linux distro, prefer privacy-first (no cloud options), or want broader GPU support via Vulkan.

Choose hyprwhspr if you're on Arch/Omarchy with Hyprland and want cloud API fallbacks when local transcription is too slow, or if you want to use NVIDIA's Parakeet model.

Both tools now offer excellent feature parity: word replacements, spoken punctuation, interactive setup, Waybar integration, and audio feedback. The main differences are architecture (Rust vs Python) and cloud capabilities.