GUIDE

An eSpeak NG alternative with natural voices

eSpeak NG is everywhere: in Linux screen readers, robots, scripts and tiny devices. It is compact, fast and speaks more than 100 languages, and it sounds robotic, which its own documentation says plainly.

The free SuperClonica engine runs the neural Supertonic 3 model offline on an ordinary processor, with natural voices in 31 languages. This guide compares the two honestly and shows how to switch a script from one to the other.

Browse voices Install the SuperClonica engine

eSpeak NG in short

For screen readers, very small devices and rare languages, it is hard to beat.

  • A compact formant synthesizer for more than 100 languages and accents.
  • Clear and usable at high speeds, but, as its README says, not as natural as synthesizers based on recordings.
  • Partial SSML support and an MBROLA backend.
  • GPL-3.0 license, actively maintained.

Hear the difference free

On the page of any voice, type the sentence your script speaks, 30 to 100 characters, choose the language and press Voice it. In a few seconds you hear the difference, with no account.

What changes with SuperClonica

Choose by the job: size and speed, or how pleasant the voice is to listen to.

  • Natural neural voices: ten stock voices and more than 170 catalog voices.
  • 31 languages, fewer than eSpeak NG; Chinese is not among them.
  • Bigger: the model is about 385 MB, yet a short phrase takes 0.43 s on four threads of a Ryzen 7 5800X in our tests.
  • Apache 2.0 for the engine and OpenRAIL-M for the model.

Switch your commands

Install the engine once: python -m pip install superclonica-cpu. A command such as espeak-ng "Backup finished" -w hello.wav becomes superclonica --text "Backup finished" --voice F1 --lang en --output hello.wav. In Python, the engine’s synthesize method replaces calls to eSpeak NG or pyttsx3, and for programs in other languages the Docker image offers the same voices over HTTP. Create the engine once and reuse it: loading takes a few seconds, each phrase a fraction of a second.

When eSpeak NG stays the better choice

Keep eSpeak NG for screen readers that need an instant answer at very high speed, for devices with very little memory or storage, for languages Supertonic 3 does not speak, and wherever a robotic voice is perfectly fine. Use SuperClonica where people listen for longer: videos, announcements, audiobooks, assistants and games.