xavierk/voculinux
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases 3 Wiki Activity
Files
f456160c455d460b510fa8b07073603db7e77fd8
voculinux/docs/USER_GUIDE.md
T
Jatin K MalikandGitHub e0ac5e409c feat(whisper): add guided whisper.cpp model variants (#465) 
* feat(whisper): refine whisper.cpp model selection

* refactor(whisper): exclude diarization model from picker

* docs(whisper): document guided model variants

* fix(ui): cap settings dialog height at 600px to prevent excessive height on high-res displays
2026-06-11 22:22:43 -07:00

6.7 KiB
Raw Blame History

User Guide

This guide explains how to use Vocalinux effectively.

Getting Started

After installing Vocalinux (see the Installation Guide), you can start the application from the terminal and optionally enable start-on-login.

Start on Login (Autostart)

Vocalinux supports login autostart using the standard Linux desktop-session mechanism.

  • Where to enable it:
    • First-run welcome dialog
    • Tray menu: Start on Login
    • Settings dialog: Start on Login
  • What Vocalinux creates:
    • vocalinux.desktop in $XDG_CONFIG_HOME/autostart/ or ~/.config/autostart/
  • How it starts:
    • As a regular user GUI app in your desktop session (--start-minimized)
  • What it does not do:
    • It does not create a systemd service/unit for autostart

Desktop Compatibility Notes

  • Works on most mainstream desktop environments (GNOME, KDE, Xfce, Cinnamon, MATE, LXQt)
  • On minimal/custom window managers, autostart may require an autostart manager or desktop-specific startup hook

Basic Usage

Starting and Stopping Voice Typing

  1. Launch the application: Run vocalinux in a terminal or launch it from your application menu
  2. Find the tray icon: Look for the microphone icon in your system tray
  3. Start voice typing: Click the tray icon and select "Start Voice Typing", or use your configured shortcut mode (toggle or push-to-talk)
  4. Speak clearly: As you speak, your words will be transcribed into the currently focused application
  5. Stop voice typing: Click the tray icon and select "Stop Voice Typing" when you're done, double-tap again in toggle mode, or release the key in push-to-talk mode

Understanding the Status Icons

  • Microphone off (gray): Voice typing is inactive
  • Microphone on (blue): Voice typing is active and listening
  • Microphone processing (orange): Voice typing is processing your speech

Voice Commands

Vocalinux supports several commands that you can speak to control formatting:

Command Action
"new line" or "new paragraph" Inserts a line break
"period" or "full stop" Types a period (.)
"comma" Types a comma (,)
"question mark" Types a question mark (?)
"exclamation point" or "exclamation mark" Types an exclamation point (!)
"semicolon" Types a semicolon (;)
"colon" Types a colon (:)
"delete that" or "scratch that" Deletes the last sentence
"capitalize" or "uppercase" Capitalizes the next word
"all caps" Makes the next word ALL CAPS

Tips for Better Recognition

  1. Use a good microphone: A quality microphone significantly improves recognition accuracy
  2. Speak clearly: Enunciate your words clearly but naturally
  3. Moderate pace: Don't speak too quickly or too slowly
  4. Quiet environment: Minimize background noise when possible
  5. Learn commands: Familiarize yourself with voice commands for punctuation and formatting
  6. Use GPU acceleration: If you have a GPU (AMD, Intel, or NVIDIA), whisper.cpp will automatically use it for faster transcription
  7. Choose the right model:
    • For real-time dictation: Use tiny or base (fastest)
    • For better accuracy: Use small, medium, or large
    • For English-only dictation: Choose an .en specialization
    • For lower-memory systems: Choose a quantized specialization such as q5_0 or q5_1
  8. Check debug logs: Run vocalinux --debug to see which backend is being used (Vulkan, CUDA, or CPU)

Customization

Keyboard Shortcut

Vocalinux supports two shortcut modes for controlling voice typing:

  • Toggle mode (default): Double-tap the shortcut key (Ctrl by default) to start/stop voice typing
  • Push-to-talk mode: Hold the configured shortcut key to speak, then release to stop
  • Configure mode and key in Settings -> Shortcuts

Model Settings

You can change the speech recognition engine and model for better accuracy or faster performance:

Choosing Your Engine

Vocalinux now offers three speech recognition engines:

  1. whisper.cpp ⭐ (Default) - High-performance C++ engine

    • Fastest installation (~1-2 min)
    • Works with AMD, Intel, NVIDIA GPUs via Vulkan
    • True multi-threading (no Python GIL)
    • Best for most users

  2. Whisper (OpenAI) - PyTorch-based engine

    • NVIDIA GPU only (requires CUDA)
    • Larger download (~2.3GB with PyTorch)
    • Installation takes ~5-10 min
    • Use if you specifically need PyTorch features

  3. VOSK - Lightweight engine

    • Smallest footprint (~40MB)
    • CPU only
    • Great for older systems or minimal resource usage

Changing Engine and Model

  1. Open settings from the tray icon menu (right-click)
  2. Go to the "Speech Engine" tab
  3. Select your Speech Engine:
    • whisper_cpp (recommended)
    • whisper
    • vosk
  4. Select your Model Size:
    • tiny (~74MB) - Fastest, good for real-time dictation
    • base (~141MB) - Good balance
    • small (~465MB) - Better accuracy
    • medium (~1.5GB) - High accuracy
    • large (~3.0GB) - Best accuracy, slower
  5. For whisper.cpp, select a Specialization:
    • Standard multilingual - Best default for auto-detect or non-English dictation
    • English-only - Choose when you dictate only in English
    • Quantized - Lower memory and smaller downloads with a possible accuracy tradeoff
    • Turbo - Faster large-v3 option with strong accuracy
    • Legacy large - Use only if you specifically need an older large model version

English-only whisper.cpp specializations limit the language selector to English.

When to Use Each Model

For real-time dictation: Use tiny or base - they're fast enough to keep up with your speech.

For transcription: Use small or medium - better accuracy for recorded audio.

For maximum accuracy: Use large - best results but requires more RAM and GPU power.

For lower-memory systems: Use a quantized whisper.cpp specialization such as Q5 or Q8.

For English-only dictation: Use an English-only specialization and keep the language set to English.

GPU Acceleration

whisper.cpp automatically uses GPU acceleration when available:

  • Vulkan (AMD, Intel, NVIDIA) - Automatically detected and used
  • CUDA (NVIDIA only) - Fallback if Vulkan not available
  • CPU - Always works as fallback

To check which backend is being used, look for these log messages when starting Vocalinux:

[INFO] whisper.cpp using Vulkan GPU backend: AMD Radeon RX 6800
[INFO] whisper.cpp configured with n_threads=16

Troubleshooting

If you encounter issues, check the Installation Guide troubleshooting section or run the application with debug logging:

vocalinux --debug

Check the logs for error messages and possible solutions.

Reference in New Issue View Git Blame Copy Permalink