Filmina vs NotebookLM

NotebookLM turns your PDF into a podcast where two hosts talk about it. Filmina explains the slide in front of you, while you look at it. Both are audio, but they are not for the same thing.

NotebookLM is the Google tool that ran past everything else on campus. You upload the sources, press a button and a while later you have a nine or ten minute audio where two hosts chat about your material, ask each other questions and explain it as if it were a radio show. It is impressive the first time, and it is free.

Then comes the moment of sitting down to study, and that is where the problem shows up.

NotebookLM tells you what the PDF is about. Filmina explains what slide 22 says, while you are looking at slide 22.

What NotebookLM does well

As a first pass over a subject you have never touched, it is excellent. If you haven't tried it yet, try it.

Where it stops being enough

There is no slide in front of you

The podcast exists on its own. You hear «the key concept here is conditional independence» and there is no way to know that came from slide 12, or to look at the formula while it is being explained to you. When the exam comes, what you will have in front of you is the PDF, not the audio, and the bridge between the two is yours to build.

Filmina doesn't exist without the PDF: what you hear is the slide you are looking at, and while it speaks a frame surrounds the block it is talking about. When you open that PDF again in finals week, you recognise the page because you have already been through it.

It picks what to tell, and doesn't tell you what it left out

A ten-minute summary of sixty slides is, necessarily, a selection. The two hosts grab whatever strikes them as juicy, build a narrative and leave out half. It is usually a good selection. But it is not your syllabus: it is theirs.

Filmina goes in the order of the document and covers every block. And not because we ask the model nicely: after each explanation is generated it is verified in code that no block of text on the page was left uncovered, and if one is missing it is requested again on its own. Coverage is checked, not trusted.

You don't get to choose the depth

The audio comes at whatever length it comes. Filmina has three levels and they change at any moment: summary for the quick pass, full for everything the slide says, extended when you also want examples, the why and the usual mistakes. The same slide, three depths, without uploading anything again.

It has quotas, and it needs internet

The free plan allows a few audios per day, and each one is generated on Google's servers. Filmina generates on your Mac, with no limit on anything, and the whole document is prepared behind you while you listen to the first slide.

Your notes get uploaded

For NotebookLM to work, the PDF has to be on Google. For a lot of material that makes no difference. For a thesis draft, a report from work or course material that isn't yours, maybe it does.

Filmina vs NotebookLM, point by point

Comparison as of September 2026. NotebookLM changes often: it is worth checking quotas and plans on Google's site.
What mattersFilminaNotebookLM
What you hear The explanation of the slide you are looking at Two hosts talking about your sources as a whole
Relation to the page A frame surrounds the block being explained None: the audio goes its own way
Coverage Every block, verified in code A selection, at whatever length comes out
Depth Three levels, switchable on the fly Whatever the model decides
Several sources together No: one PDF at a time Yes, dozens per notebook
Asking it things There is no chat Yes, with citations to the source
Where it runs On your Mac, offline On Google's servers
Usage limits None A few audios per day on the free plan
Pricing One payment of US$ 39 at launch, US$ 19 for students Free, with paid plans for more quota

Two different uses of the same PDF

The honest comparison is not which one is better, but which moment each one is for.

NotebookLM, for the first pass

You don't know what the course is about, you are on your way to class and you want to arrive with a general idea. Or you have five different sources and you need to see them together.

Filmina, for the pass that counts

You already know what it is about, you are sitting with the slide in front of you and what you need is someone to explain that slide, skipping nothing and pointing at the part they are talking about.

Using them together works well: the podcast on the commute, Filmina at the desk.

My take

Ignacio Prados @ignaprados

I use NotebookLM a lot. I turn the lecture into a podcast and listen to it while I do something else: driving, travelling, or with my hands busy on anything else. For revising it is great.

What does not happen is understanding a lecture by listening to it straight through. For that I need to see the slides and hear the explanation at the same time, and that is why I made Filmina.

What NotebookLM has and Filmina doesn't


Frequently asked questions

Is Filmina an alternative to NotebookLM?

For studying a PDF with the screen in front of you, yes. For building a notebook with ten sources and asking it questions, no: NotebookLM does that and Filmina doesn't.

Does NotebookLM make audio in languages other than English?

Yes, it is no longer English only. Some options, like choosing the length of the audio, are still English only.

How many audios can I make per day?

On the free plan it is a few, and it goes up with Google's paid plans. In Filmina there is no quota: the limit is your Mac.

Does Filmina also make a podcast out of the PDF?

No. It is one voice explaining the slide you are looking at, in the order of the document, marking on the page what it is talking about.

Are my notes uploaded to a server?

In NotebookLM yes, because the processing is Google's. In Filmina no: the model and the voice are installed on your Mac and the file never leaves your folder.


Keep comparing

Start with
just one slide

Drop in that PDF you've been putting off and hear it explained. Thirty seconds is enough to know whether it's for you.

No account · macOS 13 or later