NotebookLM's June 2026 update gave every notebook its own cloud computer. It gave Audio Overviews nothing at all.
That gap is why people keep working around it. Across the more than 162,000 sources ever uploaded to Jellypod, over 175 are explicitly titled or described as a NotebookLM Audio Overview, someone taking Google's own output and bringing it somewhere they could edit it, host it, and put their own hosts on it. Google's June 8 update, the largest NotebookLM release since Audio Overviews launched, was a real leap for research. For anyone who wanted the podcast side to grow up alongside it, this is what actually shipped, and what still didn't.
What actually changed on June 8
The update rolled out on the web starting June 8, 2026, and it is substantial:
- A newer model. NotebookLM now runs on Gemini 3.5, with an emphasis on both output quality and showing its work.
- Code execution. Every notebook gets a sandboxed environment that can write and run code in the background while you ask questions in plain English, useful for pulling structure out of messy data sources.
- Visible reasoning. When NotebookLM works through a question, the chat now shows the steps it took, not just the final answer.
- Agentic source discovery. You can start a notebook from a loose idea instead of a pile of links, and it goes and finds sources to build the repository for you.
- New export formats. Charts as PNG or SVG, plus PDF, Word, Markdown, Excel, PowerPoint, CSV, JSON, and image exports.
Access is limited for now to Google AI Ultra subscribers and Workspace Business customers with AI Ultra or Expanded Access, the same pattern NotebookLM followed when it rolled the million-token context window out to paid tiers first and free users later.
Read that list again and notice what's missing: nothing on it is about audio.
Did Audio Overviews get better?
No. Every documented piece of the June update is about research, structure, and export, not about the two-host conversation Audio Overviews generate. The feature that made NotebookLM famous is functionally the same tool it was in 2025:
- Still one Audio Overview per generation, not an editable script.
- Still two fixed AI hosts you cannot name, brief, or keep consistent across a series.
- Still no RSS feed and no path to Spotify, Apple Podcasts, or YouTube.
- Still nothing to catch a misstated fact before it ships as audio, since there is no script to review.
None of the new export formats include an audio file or a podcast feed. You can now export a chart from your notebook as an SVG. You still cannot export the Audio Overview as anything other than what NotebookLM decided to generate.
Why hundreds of people already route around this
This is not a hypothetical gap. It shows up in what people actually upload. Jellypod shipped Import from NotebookLM two weeks before Google's update, specifically because so many people were already generating an Audio Overview, deciding they liked the content but not the fixed format, and then hunting for a way to turn it into something they controlled.
The feature takes any NotebookLM Audio Overview, transcribes it, separates the two speakers, and hands you back an editable script voiced by hosts you choose, instead of the two voices NotebookLM assigned. It is the difference between a one-off you listen to once and a first episode of a real show.
| NotebookLM (post-June 2026) | Jellypod | |
|---|---|---|
| Script you can edit before audio generates | No | Yes |
| Hosts you name and keep consistent across episodes | No | Yes |
| Voice cloning | No | Yes |
| RSS feed and distribution to Spotify, Apple, YouTube | No | Yes |
| Import an existing NotebookLM Audio Overview | N/A | Yes, via Import from NotebookLM |
| Free tier | Limited, new features gated to paid Ultra/Workspace | Yes |
How to turn a NotebookLM Audio Overview into a real episode
- Generate the Audio Overview in NotebookLMUpload your sources and generate the Audio Overview the way you always have. None of this changes with the update.
- Open Import from NotebookLM in JellypodIn a podcast, click the upload icon next to Create New Episode and choose Import from NotebookLM.
- Pick your hosts and drop in the audioChoose the two hosts you want (your own cloned voice, if you have one), then upload the Audio Overview file, up to 30 minutes long.
- Edit the transcribed scriptJellypod transcribes the audio, separates the speakers, and drafts an editable script. Fix anything before you regenerate it in your hosts' voices.
- Publish to a real feedPush the episode to your podcast's RSS feed so it shows up on Spotify, Apple Podcasts, and YouTube, something no Audio Overview can do on its own.
See Importing a NotebookLM Audio Overview for the full walkthrough, or the full NotebookLM-to-published-podcast workflow if you're starting from scratch instead of an existing Audio Overview.
Frequently asked questions
Did NotebookLM's June 2026 update change Audio Overviews?
No. The update added a newer model, code execution, visible reasoning, agentic source discovery, and new export formats, all aimed at research and analysis. Audio Overviews still generate as a single two-host conversation with no script editing, no custom hosts, and no export to an audio file or podcast feed.
Is NotebookLM's June 2026 update available to everyone?
Not yet. Access launched limited to Google AI Ultra subscribers and Workspace Business customers with AI Ultra or Expanded Access. Google has said access will expand over time, following the same rollout pattern as past NotebookLM features.
What is "NotebookLM 2.0"?
It's an informal name people are using for the June 8, 2026 update, not an official version number from Google. The update is a substantial upgrade to NotebookLM's research capabilities (model, code execution, reasoning visibility, source discovery, exports), bundled together rather than shipped as separate incremental releases.
Can I still turn a NotebookLM Audio Overview into a real, distributed podcast?
Yes, but not inside NotebookLM itself. Import from NotebookLM in Jellypod takes the Audio Overview file, transcribes and separates the speakers, and rebuilds it as an editable script you can voice with your own hosts and publish to a real RSS feed.
What's the difference between NotebookLM and Jellypod now?
NotebookLM is the faster path to a one-off audio summary of your sources, and after the June update it's also a genuinely stronger research tool. Jellypod is built for the version of that idea you want to keep: named hosts, an editable script, voice cloning, and a real podcast feed that publishes to Spotify, Apple Podcasts, and YouTube. See the full comparison of NotebookLM alternatives if you're weighing more than these two.
The short version
Google's June 2026 update made NotebookLM a better research tool and left the podcast feature exactly where it was. If an editable script, consistent hosts, and real distribution are what you actually wanted from an Audio Overview, over 175 people have already made the same move: generate in NotebookLM, then bring it into Jellypod to make it a show.