I tried the Gemini Notebook newsletter workaround for a week. Here is the honest math.

A friend sent me the workaround the way people send diet tips: like it was too good to be true. Paste your newsletters into Google's Gemini Notebook (formerly NotebookLM), and it turns them into a podcast. For free. So I decided to try it out and make an experiment out of it. I ran it for a week and timed every minute.

I had heard so often: you never have to pay for newsletter-to-podcast, Google does it for free. Hearsay deserves measurement, so for seven days I ran the free route on three daily newsletters (1440, The Rundown AI, and Morning Brew): one notebook, one Audio Overview a day, hands-on minutes timed.

The short answer

The free route works. Gemini Notebook turns three pasted newsletters into a good two-host conversation, and it costs nothing. But nothing arrives on its own: no schedule, no episode waiting in the podcast app you already use. The week cost about five minutes of hands-on work a morning, plus a ten-minute wait for each episode to render. As a one-off, it is great. As a daily habit, it's much more complicated.

The method, step by step

Google's official workflow is simple: create a notebook, add a source, click Generate. What each step costs with daily newsletters as the sources:

1. Open one notebook. I created one for the week and reused it, one Audio Overview per day.
2. Get the text out of your email. The unglamorous part: no email import, no inbox connection, no forwarding address in Google's list of supported inputs. Per newsletter: open, copy, paste.
3. Click Generate and wait. Google says large notebooks can take several minutes to generate. Then nothing happens until you come back.
4. Go listen to your podcast. Play it in the app, share a link, or download the file; it never appears in your podcast app.

Here is Google's own sample of the output: their launch-post Audio Overview, two hosts bantering through source material. The format is actually quite good; everything around it is manual though.

The week, day by day

Day 1 (Sep 26): about 7 minutes. A sign-in detour first; the notebook was created at 19:25 CEST, three sources pasted by 19:35. I condensed the texts to fit the paste box: about two and a half minutes each. Day 2 (Sep 27): 11 minutes. Clean sign-in; same three newsletters in full, no caps, no condensation.

Days 3 and 4 (Sep 28 and 29): a few minutes each. By now it was the minimum: open, paste three texts, Generate, about four and a half minutes first paste to click (a floor; the timer undercounted). On day 4 none of the three had published yet, so I ran the Sep 28 editions.

Days 5 through 7 (Sep 30 to Oct 2): about 3 minutes each. Uncheck yesterday's sources, paste, rename, Generate. Nothing changed; it just stopped being new.

Then the wait: 8.5 minutes from clicking Generate to the playable episode, realistically it was closer to 9 to 13. The audio itself ran 24 and a half minutes.

What it cannot do

• No schedule. Google documents only manual, per-notebook generation. Nothing runs overnight; there is no wake-up-to-it edition, ever.
• No podcast feed. Google documents no RSS feed or podcast-app subscription for Audio Overviews. You play it in their app, share a link, or download the file; it never lands where your other podcasts live.
• The 3-per-day cap. The free tier allows 3 Audio Overviews per day, up to 50 sources per notebook. Plenty for an occasional newsletter; a full daily inbox turns it into rationing.
• The daily assembly. Every newsletter takes manual work, every day, by you, and nothing runs unless you run it: miss one busy morning and the pile is back.

So who is it for?

For the occasional long read turned into a conversation, the free route is genuinely good; the math works at a small scale. A daily inbox is a different problem: per-item manual work, a 3-per-day cap, and no feed. For the full picture, read the breakdown of the three real ways to listen to your newsletters as a podcast. An explainer on private podcast feeds is coming soon.

I liked the workaround. It is a good tool for a single long read. It is not how I listen to my newsletters every morning.

Where the facts come from

• Free-tier limits (3 Audio Overviews per day, 50 sources per notebook): Google's usage limits page, accessed 2026-09-26.
• Supported inputs (pasted text, web URLs, docs, YouTube; no paywalled pages; 500,000 words or 200 MB per source max): Google's sources help page, accessed 2026-09-26.
• The official workflow (create a notebook, add a source, click Generate): Google's Audio Overview announcement, accessed 2026-09-26.
• The rename (NotebookLM became Gemini Notebook on July 16, 2026; same product): Google's announcement, accessed 2026-09-26.
• Timing (as logged above): measured by me, Sep 26 to Oct 2. Days 1 and 2 are wall-clock; days 3 to 7 are estimates calibrated from a timed step-by-step breakdown, because per-click timing undercounted real time. The render figure is one timed sample (Oct 2; 24:32 of audio).