Skip to main content

​
Command

voxtype config

​
Description

Displays the current configuration settings loaded from config.toml, including all sections: hotkey, audio, engine, whisper, parakeet, moonshine, sensevoice, output, status, and integration. Also shows the output chain status (wtype, ydotool, clipboard availability) and model directories. This command is useful for:
  • Debugging: Verify which settings are active
  • Validation: Check if CLI overrides or environment variables are taking effect
  • Troubleshooting: Confirm model paths, output drivers, and backend configuration

​
Output Format

The output is structured into sections matching the config file:

​
Hotkey Section

[hotkey]
  key = "SCROLLLOCK"
  modifiers = []
  mode = PushToTalk
Shows the hotkey configuration, including the activation mode (PushToTalk or Toggle).

​
Audio Section

[audio]
  device = "default"
  sample_rate = 16000
  max_duration_secs = 120

[audio.feedback]
  enabled = true
  theme = "default"
  volume = 0.5
Displays audio device, sample rate, maximum recording duration, and feedback sound settings.

​
Engine Section

[engine]
  engine = Whisper
Shows the active transcription engine: Whisper, Parakeet, Moonshine, SenseVoice, Paraformer, Dolphin, or Omnilingual.

​
Whisper Section

[whisper]
  model = "base"
  language = Auto
  translate = false
Displays the Whisper model name, language configuration, and translation setting. If threads is set, it appears here.

​
Parakeet Section (Experimental)

[parakeet] (EXPERIMENTAL)
  model = "parakeet-tdt-0.6b-v3"
  on_demand_loading = false
  available models: parakeet-tdt-0.6b-v3, parakeet-tdt-0.6b-v3-int8
Shows installed Parakeet models (ONNX-based engine). If no models are installed, displays (not configured).

​
Moonshine Section (Experimental)

[moonshine] (EXPERIMENTAL)
  model = "base"
  quantized = false
  on_demand_loading = false
  available models: moonshine-base, moonshine-tiny
Shows installed Moonshine models and configuration.

​
SenseVoice Section (Experimental)

[sensevoice] (EXPERIMENTAL)
  model = "small"
  language = Auto
  use_itn = true
  on_demand_loading = false
  available models: sensevoice-small
Shows installed SenseVoice models and inverse text normalization (ITN) setting.

​
Output Section

[output]
  mode = Type
  fallback_to_clipboard = true
  driver_order = (default: wtype -> dotool -> ydotool -> clipboard)
  type_delay_ms = 0
  pre_type_delay_ms = 100
  restore_clipboard = false
  restore_clipboard_delay_ms = 200

[output.notification]
  on_recording_start = false
  on_recording_stop = false
  on_transcription = false
Shows output mode, driver fallback chain, typing delays, clipboard restoration settings, and notification preferences.

​
Status Section

[status]
  icon_theme = "emoji"
  (resolved icons: idle="🎤" recording="🔴" transcribing="⚙️" stopped="💤")
Displays the icon theme and the actual icons used for each state. Icon themes include: emoji, nerd-font, material, phosphor, codicons, omarchy, minimal, dots, arrows, text, or a path to a custom theme file.

​
Integration Section

[integration]
  state_file = "auto"
  (resolves to: "/run/user/1000/voxtype/state")
Shows the state file path for external integrations (Waybar, polybar). "auto" resolves to the runtime directory.

​
Output Chain Status

The command also displays which output tools are available: 
Output Chain:
  Display server:  Wayland (WAYLAND_DISPLAY=wayland-1)
  wtype:           ✓ installed (/usr/bin/wtype)
  eitype:          ✗ not installed
  ydotool:         ⚠ installed (/usr/bin/ydotool), daemon not running
  wl-copy:         ✓ installed (/usr/bin/wl-copy)

  → Text will be typed via wtype (CJK supported)
This section shows:
  • Display server: Wayland or X11
  • Tool status: Installed, available, or missing
  • Primary method: Which tool will be used for typing

​
Config and Model Paths

---
Config file: "/home/user/.config/voxtype/config.toml"
Models dir: "/home/user/.local/share/voxtype/models"

​
Examples

​
Basic Usage

# Show full configuration
voxtype config

​
Check Active Model

# Verify which Whisper model is configured
voxtype config | grep -A5 "\[whisper\]"
Output: 
[whisper]
  model = "large-v3-turbo"
  language = Auto
  translate = false

​
Check Output Chain

# See which output method will be used
voxtype config | grep -A10 "Output Chain"
Output: 
Output Chain:
  Display server:  Wayland (WAYLAND_DISPLAY=wayland-1)
  wtype:           ✓ installed (/usr/bin/wtype)
  ...
  → Text will be typed via wtype (CJK supported)

​
Verify CLI Overrides

# Check if --model flag is taking effect
voxtype --model medium config | grep "model ="
Output: 
  model = "medium"

​
Check Available Models

# List installed Parakeet models
voxtype config | grep "available models:"
Output: 
  available models: parakeet-tdt-0.6b-v3, parakeet-tdt-0.6b-v3-int8
  available models: moonshine-base, moonshine-tiny

​
Notes

  • Configuration precedence: CLI flags override environment variables, which override config file settings, which override built-in defaults.
  • Path resolution: The config command shows resolved paths (e.g., "auto" becomes the actual runtime directory path).
  • Model validation: The output shows whether configured models are actually installed.
  • Output chain: The display updates based on the current display server (Wayland vs X11) and installed tools.

​
See Also

Build docs developers (and LLMs) love

Get started for freeTalk to us