NotebookLM is an experimental tool by Google that acts as a virtual research assistant. Powered by Gemini, it provides a structure tailored for research tasks—allowing you to upload multiple documents, keep everything organised, and more. Thanks to Gemini’s new multi-modal capabilities, the possibilities have expanded significantly. You can add up to 50 sources, including documents, text, websites, and even YouTube videos.

What really caught my attention about NotebookLM is its new audio overview feature. It summarises all your content into a podcast-style audio clip with two voices.

I decided to test this feature with my master’s thesis, the largest single document I've written. To be thorough, I uploaded both the PDF (including its side-note format) and a plain text version, along with the website I created to share my work.

To put it simply—I was blown away. The quality of audio generation is improving across AI tools, whether it's through ChatGPT’s voice or tools like ElevenLabs. But what really stood out was how well it captured the content. It nailed every nuance, covering each section naturally. I don’t think I could have discussed an academic document more effectively myself!

I highly recommend giving it a listen.

Note: This article has been edited for grammar and readability by ChatGPT.