GUIDE

An alternative to StyleTTS 2 and Parler-TTS in 31 languages

StyleTTS 2 and Parler-TTS are two well-known open text-to-speech projects. The StyleTTS 2 paper reports human-level ratings on English datasets, and Parler-TTS from Hugging Face lets you describe a voice in plain words. Both are research code first: the StyleTTS 2 repository was last updated in 2024, Parler-TTS in December 2024.

The free SuperClonica engine is a pip package that runs on an ordinary processor, speaks 31 languages and comes with a catalog of ready voices. This guide compares them honestly and says when StyleTTS 2 or Parler-TTS is the better pick.

Browse voices Install the SuperClonica engine

StyleTTS 2 and Parler-TTS in short

For English research, fine-tuning experiments or voices described in words, both are serious tools.

  • StyleTTS 2: English models trained on LJSpeech and LibriTTS, MIT-licensed code, and style transfer from a reference clip with the LibriTTS model.
  • Its pretrained models come with a condition: tell listeners the speech is synthesized unless you have permission to use the voice. Inference relies on the GPL-licensed espeak-ng phonemizer.
  • Parler-TTS: Mini, 880 million parameters, and Large, 2.3 billion, in English, plus Mini Multilingual v1.1 in eight European languages, all under Apache 2.0.
  • A Parler-TTS voice is set by a text description of gender, pace and pitch, or by one of 34 named speakers.

Compare the sound free

On the page of any voice, paste a line you voiced with StyleTTS 2 or Parler-TTS, 30 to 100 characters, choose the language and press Voice it. In a few seconds you can compare on your own text, with no account.

What SuperClonica does differently

It runs on an ordinary processor: in our tests a short phrase took 0.43 s on four threads of a Ryzen 7 5800X.

  • 31 languages, including Russian, Japanese, Korean, Turkish, Hindi and Arabic; Chinese is not among them.
  • Ready voices: ten stock voices and more than 170 catalog voices with free demos.
  • Each voice is a JSON file of about 0.3 MB and speaks every language of the model.
  • One command to install, python -m pip install superclonica-cpu, or a Docker image with an HTTP API.

Describing a voice or training one

Parler-TTS takes a new description with every request, and StyleTTS 2 adapts the style of a reference clip. In SuperClonica, a voice is trained once in the studio from a clean 5 to 10 second recording, on a computer with an NVIDIA graphics card, or ordered on the exchange from an author who trains it. After that it runs on any processor and in all 31 languages, with a trace of the accent of the original recording. Clone only voices whose owners agreed.

Which one to choose

Choose StyleTTS 2 for English research and fine-tuning, and Parler-TTS when you want to describe a voice in words in English or one of its eight European languages. Choose SuperClonica for the other languages, for ready narrator and character voices, for speed on a processor, and for an install that is one pip command or one Docker image.