Readings, lecture slides, a research paper, a web article, a YouTube lecture, even a photo of your handwritten notes. Add your material and the podcast is built from it — grounded in what you actually gave it.
Choose from natural US and UK voices, male and female, and preview each with one tap before you build. Your hosts get names and personalities — not “Speaker 1” and “Speaker 2.”
The first voices start within seconds and the rest streams in while you listen. No staring at a spinner for ten or twenty minutes waiting for a finished file.
The hosts react to each other, jump in with quick questions, and every so often finish each other’s thought. It sounds like a conversation you’d actually want to listen to.
Open the podcast in a dedicated editor: change what a host says, give them a different voice, drop a segment, and re-render. NotebookLM’s script is locked — yours isn’t.
Pick an effort level and the podcast scales with it — and with how much source material you gave it. Great for a quick revision listen or a full-length episode.
The script is written from the documents you gave it, so the episode is accurate to your material instead of a generic AI ramble. Perfect for study, onboarding, or turning a report into something you can listen to.
When it sounds right, download a single MP3 — share it with a class, drop it into a podcast feed, or attach an audio version to a post.
Add a source, name your two hosts, and press build. You’ll be listening in seconds.