GUIDE

Windows text to speech: natural voices, offline, saved as WAV

Windows can read text aloud with its built-in voices, but they are made for reading the screen, not for producing audio files in a character’s voice. The free SuperClonica engine adds a neural model, Supertonic 3, to your PC: it saves speech as WAV files, runs offline on an ordinary processor and speaks 31 languages.

This guide shows how to install it from PowerShell, voice a phrase or a whole text file, turn a text file into audio by dropping it on a script, and choose voices.

Browse voices Install the SuperClonica engine

What you need

The ten built-in voices of the model, M1 to M5 and F1 to F5, work right away; catalog voices are JSON files that you pass with --voice.

  • Windows 10 or 11.
  • Python 3.10 or newer from python.org; the installer adds the py launcher.
  • About 400 MB of disk space for the Supertonic 3 model, downloaded once.
  • No graphics card: an NVIDIA GPU only makes it faster.

Install from PowerShell

Open PowerShell and install the engine with one command: py -m pip install superclonica-cpu. With an NVIDIA graphics card, install superclonica-gpu-cuda instead: pip brings CUDA and cuDNN itself. If the superclonica command is not found, call it as py -m superclonica, which always works.

Voice text from PowerShell

The model downloads on the first run; after that the engine works without the internet.

  • One phrase: py -m superclonica --text "Hello from Windows." --voice F1 --lang en --output hello.wav, then start hello.wav.
  • A whole document: save it in Notepad as a .txt file and use --text-file instead of --text; long texts are split at natural pauses and joined into one WAV.
  • Change --lang for other languages, for example de, es or ja.
  • On a busy PC, limit the processor threads with --threads 4.

A text file to audio with a double click

Create a file named read.bat with one line: py -m superclonica --text-file "%~1" --voice F1 --lang en --output "%~dpn1.wav" && start "" "%~dpn1.wav". Drag any .txt file onto read.bat: the engine voices it, saves a WAV with the same name next to the text and opens it in your player. Change the voice or the language in the line to suit your texts.

Voices, graphics cards and training

Every voice in the SuperClonica catalog has a page where you can voice a phrase of 30 to 100 characters for free online. Free voices download without an account; paid ones are bought once and stay yours. Windows is also where you can train your own voice: the SuperClonica worker for the studio runs on Windows 10 and 11 with an NVIDIA RTX 20 series card or newer, or a GTX 16.