There are two models available: FastSpeech and Tacotron, both of which are from TensorflowTTS.
Because the voice is generated in real-time, it has high performance requirements.
The FastSpeech model works fast but the generated voice has poor anthropomorphic effect. It works well on a middle-end device. The Tacotron has high performance requirements. Though the overall effect is better, the speed is very slow. Therefore it's bundled only for test and can't be used in practice.
Although APK downloads are available below to give you the choice, you should be aware that by installing that way you will not receive update notifications and it's a less secure way to download. We recommend that you install the F-Droid client and use that.Download F-Droid