In this tutorial, I walk you through exactly how to use Descript AI Voice for text-to-speech, step by step. Once you’re inside your Descript account, I start a New Project, choose an Audio Project (since we’re generating speech), name it, and then select “Write AI speech.” From there, I show you how the default AI voice works (like Cedric), how to type or paste a line of text, click “Done writing,” and let Descript generate the audio so you can play it back right away. Next, I show you how to switch voices by clicking the speaker name and browsing stock AI speakers. I demo swapping to a different voice (like Elizabeth) so you can hear how Descript regenerates the same script with a new sound automatically. Finally, I cover voice cloning: I create a new speaker, enable speech generation, record the required training statement (or upload a file), submit it, and wait for Descript to authorize the speaker. Once it’s approved, you can generate speech using your own cloned voice—perfect for quick narration or fixing lines without re-recording.