Speak Selection

Speak Selection is the Chrome Web Store / Edge marketplace extension: highlight text → right-click → hear it read aloud with adjustable voice, speed, pitch, and volume (Web Speech API).

This repo also includes Read Aloud, a desktop TTS app (Tk GUI + optional Linux Pro stack) that uses Microsoft neural voices via edge-tts (default: Andrew, US male).

Product	Where	Name
Marketplace extension	browser-extension-store/ · Chrome Web Store	Speak Selection
Windows desktop app	windows/	Read Aloud
Linux Pro app + native extension	read-aloud-gui.py, browser-extension/, native-host/	Read Aloud

Repo: https://github.com/beanwl/read-aloud
Privacy policy (Store): https://beanwl.github.io/read-aloud/privacy.html

Features

Speak Selection (Chrome / Edge store)

• Right-click selected text → Speak Selection
• Side panel for voice, speed, pitch, volume
• Built-in browser voices (no native host; works on any OS)

See STORE.md to pack and publish.

Read Aloud (Windows)

• Desktop GUI — paste clipboard, voice / speed / pitch / volume (0.25x–4x), Save Settings
• Local/fast voices (David / Zira) via Windows SAPI — Speakinia-style, offline, starts immediately
• Neural voices via edge-tts (Andrew / Jenny / etc.), streamed so speech starts before the full download finishes
• Taskbar / Start Menu shortcuts with the app icon (not the Python icon)
• Single-instance GUI — avoids two windows talking over each other

Read Aloud (Linux Pro)

• Desktop GUI — paste clipboard, voice / speed / pitch / volume (0.25x–4x), Save Settings
• Chrome extension — highlight text → right-click → Read Aloud; side panel for controls
• Warm speak daemon — streams audio so speech starts in about a second
• CLI — speak clipboard, selection, or PDF pages
• Single-instance GUI — avoids two windows talking over each other

Quick install (Windows)

Requirements: Windows 10/11, Python 3.10+

git clone https://github.com/beanwl/read-aloud.git
cd read-aloud
powershell -ExecutionPolicy Bypass -File windows\install-windows.ps1

That creates a venv, installs dependencies, builds windows\ReadAloud.exe (custom icon), and adds Desktop / Start Menu / taskbar shortcuts.

Or run the GUI directly after pip install -r requirements.txt in a venv:

.\venv\Scripts\pythonw.exe windows\read-aloud-gui-win.py

Settings are saved to %AppData%\read-aloud\settings.json.

Optional: load Speak Selection in Chrome via chrome://extensions → Developer mode → Load unpacked → browser-extension-store\.

Requirements (Linux Pro)

• Linux (tested on Linux Mint / Ubuntu)
• Python 3.10+
• paplay (PulseAudio) or mpv
• ffmpeg (for streaming playback; often in ~/.local/bin or apt)
• xclip (clipboard)
• Google Chrome or Brave (for the Pro extension)

Quick install (Linux Pro)

git clone https://github.com/beanwl/read-aloud.git
cd read-aloud

python3 -m venv venv
./venv/bin/pip install -r requirements.txt

# Optional system packages
sudo apt install xclip pulseaudio-utils poppler-utils   # poppler = PDF text extract

Desktop app

./launch-read-aloud-gui.sh

Or search the app menu for Read Aloud (if the .desktop file is installed).

Settings are saved to ~/.config/read-aloud/settings.json.

Chrome / Brave Pro extension

./install-browser-integration.sh
# or guided helper:
./enable-in-chrome.sh

Then in Chrome:

1. Open chrome://extensions
2. Turn on Developer mode
3. Load unpacked → select browser-extension/
4. Reload the extension after updates

Right-click selected text → Read Aloud. The side panel appears only when you start a read (not on every page).

Tester

./launch-read-aloud-tester.sh

Checks the speak daemon and plays a short sample.

CLI usage

Command	What it does
./speak.sh "hello"	Speak the given text
./speak-clipboard.sh	Speak clipboard
./speak-selection.sh	Speak mouse selection (primary)
./speak-pdf.sh file.pdf [page] [end]	Speak PDF text

Environment overrides:

export SPEAK_VOICE=en-US-JennyNeural   # voice id
export SPEAK_PLAYER=paplay             # audio player

How it fits together (Linux Pro)

Chrome extension ──nativeMessaging──► read-aloud-host.py
                                              │
Desktop GUI  ──unix socket──────────► speak-daemon.py ──► edge-tts ──► paplay
CLI speak.sh ──────────────────────► edge-tts (direct) ──► paplay

• speak-daemon.py — long-running process; streams TTS for fast start and continuous reading
• read-aloud-host.py — Chrome native messaging bridge to the daemon
• read-aloud-gui.py — Tk GUI; also talks to the daemon

Project layout

read-aloud/                        # GitHub repo slug (unchanged; keeps Pages privacy URL stable)
├── read-aloud-gui.py              # Linux Pro desktop app
├── read-aloud-tester.py           # Daemon / playback tester
├── speak*.sh                      # CLI helpers
├── windows/                       # Windows desktop app
│   ├── read-aloud-gui-win.py      # Tk GUI (edge-tts + Windows Media Player)
│   ├── install-windows.ps1        # venv, launcher exe, shortcuts, taskbar pin
│   ├── launch-read-aloud.vbs      # Shortcut helper
│   └── read-aloud.ico             # App / taskbar icon
├── browser-extension/             # Chrome MV3 Pro (native host / Linux) — Read Aloud
├── browser-extension-store/       # Chrome/Edge marketplace — Speak Selection
├── native-host/                   # Native messaging + speak daemon
├── store/                         # Listing assets + pack output
├── docs/                          # GitHub Pages (privacy.html)
├── STORE.md                       # How to publish Speak Selection
├── install-browser-integration.sh
├── enable-in-chrome.sh
├── ROADMAP.md
└── requirements.txt

Chrome / Edge marketplace (Speak Selection)

# Test: Load unpacked → browser-extension-store/
./pack-store-extension.sh   # creates store/dist/*.zip for upload

See STORE.md for Chrome Web Store and Edge Add-ons submission steps.

Privacy policy: https://beanwl.github.io/read-aloud/privacy.html (docs/privacy.html).

The Pro extension in browser-extension/ needs a local native host and is not Store-ready.

Troubleshooting (Windows)

Problem	Fix
Taskbar shows Python icon	Re-run windows\install-windows.ps1, then launch from the Read Aloud shortcut (not raw pythonw)
“No pyvenv.cfg file” popup	Re-run windows\install-windows.ps1
Taskbar says “Can't open this item”	Click Yes to remove the dead pin, then right-click Desktop Read Aloud → Pin to taskbar
No sound	Check Windows volume; confirm the status line leaves “Generating speech…”
Double voices	Close the other Read Aloud window — only one instance is allowed
python not found	Install Python 3.10+ and tick Add python.exe to PATH

Troubleshooting (Linux Pro)

Problem	Fix
Stuck on “Reading…”	Run the tester; restart daemon: pkill -f speak-daemon.py then open the GUI again
Extension menu missing	Reload unpacked extension; refresh the page
Double voices	Close extra GUI windows; only one instance is allowed
Slow start	Ensure speak daemon is running (~/.cache/read-aloud/speak.sock)
No sound	Check paplay / volume; confirm ffmpeg is on PATH

License

Use freely for personal projects. Marketplace voices are the browser’s built-in Web Speech voices. Desktop (Windows / Linux Pro) voices are provided by Microsoft’s online TTS via edge-tts — respect their terms of use.