Search for a command to run...
Neural TTS for Linux at $0 cost
No comments yet. Be the first to comment.
After trying many journaling apps, I settled on Diarium for its rich, calendar-driven features like photo, audio, and mood attachments. However, I was frustrated when I switched to Linux and found it
A Self-Hosted Health Records Keeper
How the SDD Workflow plugin for Claude Code turns vague ideas into shipping software — with traceability built in. If you've ever shipped a feature that didn't match what was asked for, or merged code
If you use Obsidian for note-taking, you know how tedious capturing content from applications on your desktop: copy, switch windows, paste, format. I wanted something faster: highlight text, hit a hot
If you prefer listening to text, Dawnstar ReadAloud makes it effortless on Linux. Many TTS options today sound robotic, are paid, or require copying text into a browser. I wanted a faster, more natural workflow: highlight any text, press a hotkey, and hear neural voices instantly. Here's how I built it
A simple TTS system that:
Uses neural voices (not robotic synthesis)
Works with global keyboard shortcuts in any app
Caches audio for instant replay of repeated phrases
Handles files, PDFs, EPUBs, and URLs
Works on both X11 and Wayland
A Python CLI that leverages Microsoft Edge TTS (Azure neural voices, free) with MD5-based caching and desktop integration.
Text Input: Accepts direct text, files, URLs, or the clipboard via xclip (X11) or wl-paste (Wayland)
Chunking: Splits text at sentence boundaries (. ! ?) for natural speech rhythm
TTS Engine: Primary Edge TTS, with gTTS and eSpeak as fallbacks
Caching: Stores audio in ~/.cache/tts_app/ keyed by md5(text + lang + speed)
Playback: Auto-detects player (mpg123 → paplay → cvlc → ffplay)
#!/bin/bash
# Install Dawnstar ReadAloud
# System dependencies
sudo apt install mpg123 xclip poppler-utils
# Clone and install
git clone https://github.com/emeeran/Dawnstar-ReadAloud.git tts && cd tts
python3 -m venv .venv
./.venv/bin/pip install -e .
# Test it
./tts "Hello! This sounds surprisingly natural."
# Set up global hotkeys
python3 configure.py
The configure.py Script auto-detects your desktop and registers hotkeys. For GNOME via gsettings:
# Speak selection: Ctrl+Alt+S
gsettings set org.gnome.settings-daemon.plugins.media-keys.custom-keybinding:/org/gnome/settings-daemon/plugins/media-keys/custom-keybindings/custom0/ name 'TTS Speak'
gsettings set org.gnome.settings-daemon.plugins.media-keys.custom-keybinding:/org/gnome/settings-daemon/plugins/media-keys/custom-keybindings/custom0/ command '/home/you/.local/bin/tts-speak'
gsettings set org.gnome.settings-daemon.plugins.media-keys.custom-keybinding:/org/gnome/settings-daemon/plugins/media-keys/custom-keybindings/custom0/ binding '<Ctrl><Alt>s'
# Stop speaking: Ctrl+Alt+Q
gsettings set org.gnome.settings-daemon.plugins.media-keys.custom-keybinding:/org/gnome/settings-daemon/plugins/media-keys/custom-keybindings/custom1/ name 'TTS Stop'
gsettings set org.gnome.settings-daemon.plugins.media-keys.custom-keybinding:/org/gnome/settings-daemon/plugins/media-keys/custom-keybindings/custom1/ command '/home/you/.local/bin/tts-stop'
gsettings set org.gnome.settings-daemon.plugins.media-keys.custom-keybinding:/org/gnome/settings-daemon/plugins/media-keys/custom-keybindings/custom1/ binding '<Ctrl><Alt>q'
# Register them
gsettings set org.gnome.settings-daemon.plugins.media-keys custom-keybindings "['/org/gnome/settings-daemon/plugins/media-keys/custom-keybindings/custom0/', '/org/gnome/settings-daemon/plugins/media-keys/custom-keybindings/custom1/']"
On X11, the primary selection (highlighted text) is prioritized over clipboard—you don't need to copy, just highlight.
Highlight any text, press Ctrl+Alt+S, and hear it spoken:
# Cache stats after using the hotkey
./tts --cache-stats
Cache Statistics:
Files: 12
Size: 4.2 MB / 500 MB
Location: /home/you/.cache/tts_app/
First playback takes ~1 second (network request), subsequent plays of the same text are instant.
./tts "Direct text" # Speak directly
./tts document.txt # Read a file
./tts book.epub # Read EPUB (skips front matter)
./tts https://example.com # Read a webpage
cat notes.txt | ./tts - # Read from stdin
./tts -l en-uk -s slow file.txt # British English, slow speed
Python 3.12+
mpg123 (recommended audio player)
xclip or wl-clipboard (for keyboard shortcuts)
poppler-utils (PDF support, optional)
# Ubuntu/Debian
sudo apt install mpg123 xclip poppler-utils
The best TTS system is the one you actually use. By reducing friction to a single hotkey, I find myself listening more—documentation while cooking, proofreading by ear, articles during walks. The neural voices make it pleasant rather than tolerable.
The modular architecture makes it easy to extend: add languages in core/constants.py, swap backends in core/engine.py. Full documentation at USER_MANUAL.md.
Happy listening!