Turn a rough transcript into clean, readable text: punctuation, speaker labels and structure. It does not take audio files — you paste the text.