GUIDE

Raspberry Pi text to speech with natural voices

A Raspberry Pi that talks turns a hobby build into something the whole family uses: a doorbell that greets guests, a weather report every morning, a plant monitor that asks for water, a quiz box for kids. The usual Pi voices, espeak-ng and pico2wave, are easy to install but sound robotic.

The free SuperClonica engine gives a Pi natural neural voices in 31 languages. This guide shows three ways to use it, from files voiced in advance to the engine running on the Pi itself, and says honestly which one fits which project.

Browse voices Install the SuperClonica engine

Three ways to make a Pi talk

If your project only ever says a few fixed phrases, pre-voiced files are all you need.

  • Pre-voiced files: voice the phrases on your computer, copy them to the Pi and play them with aplay. The fastest and most reliable way.
  • A server on your network: a PC or home server runs the Docker image, the Pi sends text and plays the answer. New text at any time.
  • The engine on the Pi: the Python package installs on a Pi 4 or Pi 5 with the 64-bit system. Slower, but fully offline on one board.
  • The voices speak 31 languages; Chinese is not supported yet.

Hear the voices free online

On the page of any voice, type a phrase your Pi will say, 30 to 100 characters, choose the language and press Voice it. In a few seconds you hear how your project will sound and can download the file, with no account.

Ask a server on your network

Keep the server on your home network or behind a VPN, and never open it to the internet without a key.

  • On a PC or home server with Docker, start the server: docker run -d -p 8000:8000 -v superclonica-models:/models superclonica/superclonica:cpu. The image is built for x86-64 processors, so it does not run on the Pi itself.
  • On the Pi, ask it and play the answer: curl -s -X POST http://192.168.1.20:8000/synthesize -H "Content-Type: application/json" -d '{"text": "The plants need water.", "voice": "F1", "lang": "en"}' -o /tmp/say.wav && aplay /tmp/say.wav
  • Set SUPERCLONICA_API_KEY on the server and add -H "Authorization: Bearer YOUR-KEY" to the request.
  • A short phrase takes well under a second on an ordinary desktop processor, plus the time on your network.

Run the engine on the Pi itself

On a 64-bit Raspberry Pi OS, create a virtual environment first, because the system Python refuses pip installs: python3 -m venv ~/tts, then ~/tts/bin/pip install superclonica-cpu. Then voice and play a phrase: ~/tts/bin/superclonica --text "Good morning! It is time for breakfast." --voice F1 --lang en --output hello.wav && aplay hello.wav. The first run downloads the model, about 385 MB. Every library the engine needs has a build for the Pi’s 64-bit ARM processor, but we have not timed it on a Pi: expect it to be several times slower than a desktop, and voice repeated phrases once and keep the files.

Projects that talk

A cron job can voice the weather or the day’s calendar every morning, a motion sensor can greet whoever comes in, and a button can play a random joke for the kids. For a project you sell or show publicly, use a voice you have the rights to: the ten stock voices of the model, M1 to M5 and F1 to F5, or your own voice trained in the SuperClonica studio from a recording of a speaker who agreed. Character voices from the catalog are fan-made and fit personal builds.