Draft phrases, not yet checked by a native speaker. The list works offline; hearing a phrase needs a connection.
0
Settings
Audio is processed in memory on our server and never stored. This is an internal test build: speech recognition, translation and the Uzbek voice are early versions.