🧠

Neural TTS

Piper uses ONNX neural models for natural, human-like voice synthesis.

🔒

100% Offline

All processing is local. No API keys, no internet connection needed during use.

🎮

Ren'Py Compatible

Works transparently with Ren'Py's built-in self-voicing (press V).

🌍

Multiple Voices

Choose from US and UK English voices. Switch with one line change.

⚡

Fast

Low latency synthesis — speech starts within milliseconds of each line.

♿

Accessibility

Makes visual novels accessible to visually impaired players on Linux.

How It Works

Ren'Py calls espeak -v <voice> -a <amp> <text> whenever self-voicing is active. This project installs a wrapper script named espeak that intercepts those calls and redirects speech synthesis to Piper TTS instead. The audio is played back immediately via aplay.

Ren'Py never knows the difference — it still calls espeak, but now it sounds natural.

Installation

Install dependencies

sudo apt install python3 python3-pip alsa-utils

Install piper-tts

pip install --user piper-tts

Or from a locally downloaded wheel: pip install --user /path/to/piper_tts-1.4.2-*.whl

Create model directory and download a voice

mkdir -p $HOME/.local/share/piper/models

# List all available voices:
python3 -m piper.download_voices

# Download UK English (jenny_dioco):
python3 -m piper.download_voices en_GB-jenny_dioco-medium \
    --data-dir $HOME/.local/share/piper/models

Install the wrapper script

# Clone this repo (or download the script manually)
git clone https://github.com/zirize/piper-renpy-selfvoicing.git
mkdir -p $HOME/.local/share/piper/scripts
cp piper-renpy-selfvoicing/piper_espeak_wrapper.sh $HOME/.local/share/piper/scripts/espeak
chmod +x $HOME/.local/share/piper/scripts/espeak
ln -sf $HOME/.local/share/piper/scripts/espeak $HOME/.local/bin/espeak

Test

espeak "Piper installation complete."

If you hear the voice, you're done! Launch a Ren'Py game and press V to enable self-voicing.

💡 $HOME/.local/bin is in PATH by default on Ubuntu 21.04+, Fedora, Arch, and most modern distros. If espeak is not found, run export PATH="$HOME/.local/bin:$PATH" or add it to your ~/.bashrc.

Available Voice Models

Browse all voices at rhasspy.github.io/piper-samples. Some recommended options:

Model ID	Language	Quality
`en_US-hfc_female-medium`	US English	Medium (default)
`en_US-amy-medium`	US English	Medium
`en_GB-jenny_dioco-medium`	UK English	Medium
`en_GB-cori-high`	UK English	High
`en_GB-alba-medium`	UK English (Scottish)	Medium
`en_GB-semaine-medium`	UK English	Medium

To switch voices, edit MODEL= at the top of ~/.local/share/piper/scripts/espeak.

Optional: Auto-Advance

The wrapper can automatically press Space after each dialogue line finishes speaking, so you don't have to click — great for listening to a whole scene hands-free.

sudo apt install xdotool
RENPY_MODE=1 ./YourGame

Repository Files

File	Description
`piper_espeak_wrapper.sh`	Main espeak wrapper (copy to `~/.local/share/piper/scripts/espeak`)
`install.sh`	Automated installation script
`guide.txt`	Detailed step-by-step guide
`note.txt`	Quick reference for voice models