BigLinux TTS

Complete text-to-speech solution with graphical configuration for BigLinux

Overview

BigLinux TTS is a native GTK4/libadwaita application for BigLinux that provides a complete text-to-speech system with graphical configuration. It supports three TTS engines (speech-dispatcher, espeak-ng, Piper Neural TTS), automatic voice discovery, customizable keyboard shortcuts, and real-time speech parameter control.

Built with modern GNOME HIG principles, it integrates seamlessly into KDE Plasma and any desktop environment running GTK4.

Features

Multiple TTS Engines — Choose between speech-dispatcher (RHVoice, espeak), espeak-ng direct, or Piper neural TTS for high-quality offline synthesis.
Automatic Voice Discovery — Scans installed TTS engines and voices automatically. Detects RHVoice voices, espeak-ng languages, and Piper neural models.
Real-Time Voice Settings — Adjust speed, pitch, and volume with live preview. All parameters are engine-aware with proper conversion for each backend.
Test Any Voice — Built-in test field where you can type any text and hear it immediately with the selected voice and settings.
Piper Neural TTS Integration — Full support for Piper neural voice models with automatic detection of .onnx models in /usr/share/piper-voices/. One-click installation of Piper engine and voice packages.
Customizable Keyboard Shortcut — Change the global shortcut (default: Alt+V) directly from the app with a visual key capture dialog. Automatically updates KDE global shortcuts.
Pin to Taskbar — Optionally pin the speak action to the KDE Plasma taskbar for one-click text reading without keyboard shortcuts.
Read Selected Text — Select any text anywhere on your desktop and press the shortcut to read it aloud. Press again to stop.
Text Processing — Expand abbreviations, read special characters by name, strip HTML/Markdown formatting, handle URLs, and set character limits.
speech-dispatcher Python API — Uses the native speechd.SSIPClient Python API for reliable speech-dispatcher communication, bypassing CLI limitations.
Wayland & X11 Support — Full clipboard integration using wl-clipboard-rs (Wayland) and xsel (X11).
Internationalization — Translated into 29 languages via gettext .po files.
Persistent Settings — All preferences saved to ~/.config/biglinux-tts/settings.json following XDG specification.
Modern UI — Adwaita design with hero section, preferences groups, expander rows, toast notifications, and responsive layout.

Screenshots

TTS Engines

speech-dispatcher (Default)

The primary TTS backend. Routes speech through the speech-dispatcher daemon, which supports multiple output modules:

Module	Description
RHVoice	High-quality multilingual voices (Letícia F123 for pt-BR, Evgeniy-Eng for English)
espeak-ng	Lightweight, supports 100+ languages
espeak	Legacy espeak engine
pico	SVOX Pico TTS
festival	University of Edinburgh's Festival system

Communication uses the Python speechd.SSIPClient API directly for reliable text delivery and callback support.

espeak-ng (Direct)

Bypasses speech-dispatcher and calls espeak-ng directly. Useful for:

Lower latency
Simpler configuration
Systems without speech-dispatcher

Supports 100+ languages and variants with configurable voice, speed (80–450 WPM), pitch (0–99), and volume.

Piper (Neural TTS)

Offline neural text-to-speech using ONNX models. Produces natural-sounding speech:

Binary: piper-tts (package: piper-tts-bin)
Voices: .onnx models in /usr/share/piper-voices/
Audio: Raw PCM output piped through aplay (22050 Hz, 16-bit)
Volume: Adjusted via sox pipeline when available
Parameters:
- --length_scale — Speed control (0.3 = fast, 1.0 = normal, 2.5 = slow)
- --noise_scale — Voice expressiveness/intonation variation
- --noise_w — Phoneme duration variation
- --sentence_silence — Pause between sentences

Auto-install dialog offers to install piper-tts-bin and language-specific voice packages via pkexec pacman.

Requirements

Dependency	Package	Required
Python	`python` (3.10+)	Yes
PyGObject	`python-gobject`	Yes
GTK 4	`gtk4`	Yes
libadwaita	`libadwaita` (1.x)	Yes
speech-dispatcher	`speech-dispatcher`	Yes
espeak-ng	`espeak-ng`	Yes
X11 clipboard	`xsel`	Yes
Wayland clipboard	`wl-clipboard-rs`	Yes
ALSA utilities	`alsa-utils`	Yes
RHVoice	`rhvoice`	Optional
Piper TTS	`piper-tts-bin`	Optional
SoX	`sox`	Optional

Installation

BigLinux / Manjaro / Arch Linux

# From the official BigLinux repository
sudo pacman -S tts-biglinux

# Optional: install RHVoice Brazilian Portuguese voice
sudo pacman -S rhvoice rhvoice-voice-leticia-f123

# Optional: install Piper neural TTS
sudo pacman -S piper-tts-bin piper-voices-pt-BR

From Source

git clone https://github.com/biglinux/tts-biglinux.git
cd tts-biglinux
PYTHONPATH=src python3 -m biglinux_tts

Building the Package (makepkg)

cd pkgbuild
makepkg -si

Project Structure

tts-biglinux/
├── src/
│   └── biglinux_tts/
│       ├── __init__.py              # Package initialization
│       ├── __main__.py              # Entry point (python3 -m biglinux_tts)
│       ├── main.py                  # CLI argument parsing, logging setup
│       ├── application.py           # Adw.Application lifecycle
│       ├── window.py                # Main window (Adw.ApplicationWindow)
│       ├── config.py                # Configuration dataclasses, constants, I/O
│       ├── services/
│       │   ├── tts_service.py       # TTS speak/stop state machine
│       │   ├── voice_manager.py     # Voice discovery across all engines
│       │   ├── text_processor.py    # Text normalization, abbreviation expansion
│       │   ├── clipboard_service.py # Clipboard access (Wayland + X11)
│       │   └── settings_service.py  # JSON settings persistence
│       ├── ui/
│       │   ├── main_view.py         # Main settings view (hero, controls, sections)
│       │   ├── components.py        # Reusable widget factories (9 factory functions)
│       │   └── style.css            # Custom CSS (hero section, animations)
│       ├── utils/
│       │   ├── async_utils.py       # Thread helpers, debouncer
│       │   └── i18n.py              # Gettext internationalization helper
│       └── resources/
│           ├── __init__.py          # CSS loader (load_css)
│           └── style.css            # Application stylesheet
├── usr/
│   ├── bin/
│   │   ├── biglinux-tts             # Launch settings GUI
│   │   └── biglinux-tts-speak       # Read selected text via shortcut
│   ├── share/
│   │   ├── applications/
│   │   │   ├── bigtts.desktop                    # KDE global shortcut (Alt+V, NoDisplay)
│   │   │   └── br.com.biglinux.tts.desktop       # Settings app launcher
│   │   ├── icons/hicolor/scalable/apps/
│   │   │   └── biglinux-tts.svg                  # Application icon
│   │   └── locale/                               # Compiled translations (.mo, .json)
│   │       ├── bg/LC_MESSAGES/
│   │       ├── cs/LC_MESSAGES/
│   │       ├── ...                               # 28 languages
│   │       └── zh/LC_MESSAGES/
├── locale/                                       # Translation source files (.po, .pot)
├── pkgbuild/
│   └── PKGBUILD                                  # Arch/BigLinux package build script
├── pyproject.toml                                # Python project metadata
└── README.md

Architecture

┌──────────────────────────────────────────────────────────────┐
│                      main.py / __main__.py                   │
│                   BigLinuxTTSApp (Adw.Application)           │
├──────────────┬──────────────────────┬────────────────────────┤
│   UI Layer   │   Service Layer      │   Configuration        │
├──────────────┼──────────────────────┼────────────────────────┤
│ window.py    │ tts_service.py       │ config.py              │
│ main_view.py │ voice_manager.py     │ settings_service.py    │
│ components.py│ text_processor.py    │                        │
│ style.css    │ clipboard_service.py │                        │
└──────────────┴──────────────────────┴────────────────────────┘
                              │
              ┌───────────────┼───────────────┐
              ▼               ▼               ▼
      speech-dispatcher   espeak-ng       piper-tts
      (speechd.SSIPClient)  (direct)     (stdin → aplay)

UI Layer — GTK4 + libadwaita widgets following GNOME HIG. Hero section with live status, preferences groups for voice settings, TTS engine selection, text processing, and advanced options. Toast notifications for feedback.
Service Layer — TTS state machine with speak/stop/cleanup lifecycle. Background voice discovery across all engines. Text normalization pipeline. Clipboard access for both Wayland and X11.
Configuration — Typed dataclasses for all settings. JSON persistence at ~/.config/biglinux-tts/settings.json. Sensible defaults with reset capability.

How It Works

Reading Selected Text

Select text anywhere on your desktop (browser, editor, terminal, etc.).
Press Alt+V (or your custom shortcut).
The selected text is captured via clipboard and spoken aloud.
Press Alt+V again to stop.

Using the Taskbar Button

Open Advanced options → Show speak button in app launcher.
Toggle the switch ON.
The TTS icon is pinned to the KDE Plasma taskbar automatically.
Click the icon to speak selected text — click again to stop.

Configuring Voice Settings

Launch the settings app: PYTHONPATH=src python3 -m biglinux_tts (or biglinux-tts when installed).
Select a TTS Engine (speech-dispatcher, espeak-ng, or Piper).
Choose a Voice from the discovered voices dropdown.
Adjust Speed, Pitch, and Volume sliders.
Type text in the test field and click Test voice to preview.
All changes are saved automatically.

Changing the Keyboard Shortcut

Open Advanced options → Keyboard shortcut.
Click Change.
Press your desired key combination (e.g., Ctrl+Shift+S).
The new shortcut is saved and applied to KDE global shortcuts immediately.

Installing Piper Neural TTS

Select Piper (Neural TTS) as the engine.
If Piper is not installed, an installation dialog appears.
Click Install to automatically download and install via pacman:
- piper-tts-bin — The Piper binary
- piper-voices-<lang> — Voice models for your language
After installation, voices are discovered automatically.

Configuration

User preferences are stored at:

~/.config/biglinux-tts/settings.json

Settings Structure

{
  "speech": {
    "backend": "speech-dispatcher",
    "output_module": "rhvoice",
    "voice_id": "Leticia-F123",
    "rate": 0,
    "pitch": 0,
    "volume": 80
  },
  "text": {
    "expand_abbreviations": true,
    "process_special_chars": false,
    "strip_formatting": true,
    "process_urls": false,
    "max_chars": 0
  },
  "shortcut": {
    "keybinding": "<Alt>v",
    "show_in_launcher": false
  }
}

Setting	Description	Default
`backend`	TTS engine: `speech-dispatcher`, `espeak-ng`, `piper`	`speech-dispatcher`
`output_module`	speech-dispatcher module: `rhvoice`, `espeak-ng`, etc.	`rhvoice`
`voice_id`	Voice identifier (engine-specific)	Auto-detected
`rate`	Speech speed (-100 to 100)	`0`
`pitch`	Voice pitch (-100 to 100)	`0`
`volume`	Speech volume (0 to 100)	`80`
`expand_abbreviations`	Replace abbreviations with full words	`true`
`process_special_chars`	Read symbols (#, @, %) by name	`false`
`strip_formatting`	Remove HTML/Markdown before reading	`true`
`process_urls`	Read URLs aloud	`false`
`max_chars`	Character limit (0 = unlimited)	`0`
`keybinding`	Global keyboard shortcut	`<Alt>v`
`show_in_launcher`	Pin speak button to taskbar	`false`

Voice Discovery

The application automatically discovers voices from multiple sources:

RHVoice

Scans /usr/share/RHVoice/voices/ for installed voice directories. Extracts language, gender, and display name from the voice configuration.

espeak-ng

Runs espeak-ng --voices and parses the output. Maps language codes to human-readable voice names with gender detection.

Piper

Recursively scans /usr/share/piper-voices/ for .onnx model files paired with .onnx.json configuration. Extracts language, speaker name, and quality level from the file path structure: {lang}/{lang_REGION}/{speaker}/{quality}/{model}.onnx.

Translation

The application uses gettext for internationalization. Translation files are located in locale/.

Supported Languages

Bulgarian, Czech, Danish, Dutch, English, Estonian, Finnish, French, German, Greek, Hebrew, Croatian, Hungarian, Icelandic, Italian, Japanese, Korean, Norwegian, Polish, Portuguese, Portuguese (Brazil), Romanian, Russian, Slovak, Swedish, Turkish, Ukrainian, Chinese.

Adding a New Translation

Copy the template: cp locale/tts-biglinux.pot locale/<lang>.po
Edit the .po file with your translations.
Compile: msgfmt locale/<lang>.po -o usr/share/locale/<lang>/LC_MESSAGES/tts-biglinux.mo
Generate JSON: Convert the .po file to JSON for runtime use.

Command-Line Usage

# Launch the settings GUI
biglinux-tts

# Read selected text (used by keyboard shortcut)
biglinux-tts-speak

# Read text from stdin
echo "Hello, world!" | tts-text

# Launch with debug logging
PYTHONPATH=src python3 -m biglinux_tts --debug

Contributing

Fork the repository.
Create a feature branch: git checkout -b feature/my-feature
Commit your changes: git commit -m 'Add my feature'
Push to the branch: git push origin feature/my-feature
Open a Pull Request.

License

This project is licensed under the GPL-3.0 License.

Credits

BigLinux Team — Distribution and packaging
GTK / GNOME — UI framework
RHVoice — High-quality multilingual TTS
espeak-ng — Lightweight TTS engine
Piper — Neural TTS by Rhasspy
speech-dispatcher — TTS middleware

Name		Name	Last commit message	Last commit date
Latest commit History 72 Commits
.github/workflows		.github/workflows
.serena		.serena
locale		locale
pkgbuild		pkgbuild
src/biglinux_tts		src/biglinux_tts
usr		usr
README.md		README.md
pyproject.toml		pyproject.toml

biglinux/tts-biglinux

Folders and files

Latest commit

History

Repository files navigation

BigLinux TTS

Overview

Features

Screenshots

TTS Engines

speech-dispatcher (Default)

espeak-ng (Direct)

Piper (Neural TTS)

Requirements

Installation

BigLinux / Manjaro / Arch Linux

From Source

Building the Package (makepkg)

Project Structure

Architecture

How It Works

Reading Selected Text

Using the Taskbar Button

Configuring Voice Settings

Changing the Keyboard Shortcut

Installing Piper Neural TTS

Configuration

Settings Structure

Voice Discovery

RHVoice

espeak-ng

Piper

Translation

Supported Languages

Adding a New Translation

Command-Line Usage

Contributing

License

Credits

About

Resources

Uh oh!

Stars

Watchers

Forks

Releases 11

Packages 0

Uh oh!

Uh oh!

Contributors

Uh oh!

Languages

Packages