It does exactly what you think it does. You can cut parts of the original file without having to edit the waveform (like you would in Audacity). Instead, select the parts directly just like you would in a text editor.

What it does not do is generate new words (ie you type a sentence and it adds that to your file as voice).