Android. Text-to-speech

2026.03.15

Hi!

Install apps

• SherpaTTS as TTS engine itself (has basic controls for reading text)
• TTS Util for more convenient TTS interface (not a TTS engine)

Also you'll need adb installed on the desktop so you can copy model files from there.

Get voice models

SherpaTTS has a list of models which you can download from HuggingFace from the app.

If that list is not enough (or returns errors, or you want to use your own model), you can sideload other models.

Choose the model from e.g. HuggingFace, then search it on the release page and download needed files (already in onnx format).

If you need to convert from Piper model see here.

Extract downloaded archive and put those files into yourModelName directory (we will copy those to the device later):

• *.onnx renamed as model.onnx
• tokens.txt
• lang - create this file with 3-letter language code on the 1st line and model name on the 2nd

Copy voice models to device

Connect the smartphone via USB. Enable USB Debugging and choose File transfer in USB connection properties on the phone.

List connected devices and make sure you got your phone there:

adb devices

Next steps should be done one by one for each model you want to copy.

Models should be copied to /sdcard/Android/data/org.woheller69.ttsengine/files/modelDir, so run:

adb push <path-to-model-files>/ /sdcard/Android/data/org.woheller69.ttsengine/files/modelDir

You can check the files via entering a shell into Android with:

adb shell

After copying, open SherpaTTS app. It will register model files and put them where it needs. Then, you can repeat same steps to copy other models.

Use via TTS Util

Open TTS Util, go to settings and select TTS Engine (SherpaTTS) and TTS Language/Voice. Tune other settings as needed.

Bye!