Filmina vs Speechify
Speechify reads the PDF out loud to you. Filmina explains it slide by slide and points at the part it's talking about. What each one does well, what they cost and which one gets you through Tuesday's exam.
If you got here looking for Speechify it's because you have a PDF that has to be studied and there aren't enough hours to sit and read it. Speechify is the best known app for listening to documents and it is very good at what it does. Filmina does something else, and this page is here so you can see which of the two things you need.
Speechify turns text into audio. Filmina turns a slide into an explanation, and while it says it, it marks on the slide which block it is talking about.
What Speechify is and what it does well
Speechify is a text-to-speech reader. You give it a PDF, an article, an email or a photo of a page and it reads it to you with synthetic voices that genuinely sound good. It started as an accessibility tool, for people with dyslexia, and that shows in how polished the craft of reading is.
What it does well, plainly:
- Natural voices in more than 60 languages, with a huge library to choose from.
- Up to 5x speed on the paid plan, which for re-reading something you already know is a lot.
- It is everywhere: iPhone, Android, Chrome extension, desktop and web. You start on the computer and carry on on the bus.
- It scans with the camera: you take a photo of a handout and it reads it.
- It reads anything, not just class notes: emails, papers, contracts, web pages.
If what you need is to listen to a long text while you travel or walk, Speechify is a great tool and this page is not going to try to convince you otherwise.
What Filmina is
Filmina is a desktop teacher. You drop in the PDF and, instead of reading you the slide, it explains it: what that block says, why it is there and how it connects to the previous one. While it speaks, a frame travels across the slide and surrounds the part it is talking about, so you don't have to guess which line it is on.
Three things that define it:
- It explains, it doesn't read. The text you hear is not the one on the slide: it is an explanation written from it, verified block by block so that nothing gets skipped.
- It points at what it is talking about. A frame around the block, which slides to the next one as the voice moves on.
- It runs entirely on your Mac. The model that writes and the voice that speaks are installed on your computer. The PDF is not uploaded to any server.
Filmina vs Speechify, point by point
| What matters | Filmina | Speechify |
|---|---|---|
| What you hear | An explanation of the slide, written from what is on it | The text on the page, read exactly as written |
| On the page | A frame surrounds the block being explained | It highlights the word or phrase being read |
| Headers and footers | Detected because they repeat, and discarded | Read out, because they are text like any other |
| Formulas | P(A|B) is said «P of A given B» | Read as they come out of the PDF |
| If something doesn't make sense | You raise the level of detail and it tells you with examples and usual mistakes | You slow the speed down and listen again |
| Where it runs | On your Mac, offline | In the cloud. Audio already generated can be downloaded to listen without signal |
| Your notes | Never leave your computer | Uploaded in order to be processed |
| Pricing | One payment of US$ 39 at launch, US$ 19 for students | Subscription of around US$ 139 a year, or US$ 29 a month |
| Where it works | macOS 13 or later | iOS, Android, Chrome, Windows, web |
| Languages | English and Spanish | More than 60 |
Why a slide resists a text-to-speech reader
A class PDF is not a text. It is the script of someone who was speaking. The six bullets on slide 22 were the lecturer's cues, not the explanation: the explanation was said out loud, and it isn't in the PDF.
So when a reader reads you that slide, it hands you back exactly the same six bullets you had already read, now in audio. The problem was not your reading speed. It was that «layer model 3, 4 and 7» says nothing if nobody tells you what those layers are.
And there is a duller part, though you notice it in the first minute: a PDF is full of things that are not content. The footer repeating the file path and the slide number. The department logo. The loose labels of a diagram, which without the drawing are three words with no subject. Filmina discards all of that before starting (headers and footers are recognised because they repeat on almost every page), and it groups a diagram's labels and tells them together, as what they are.
Underlining the word is not pointing at the idea
Speechify highlights the word it is saying, and that's fine: it helps you not lose your place while you listen. But it is a highlight that moves at the speed of the mouth.
Filmina doesn't follow the word: it follows the idea. The frame surrounds the whole block for as long as the explanation of that block lasts, which can be half a sentence of audio or twenty seconds. When you look up after thinking about something else, what you need to know is not which word it is on, but which part of the slide we are talking about. And the frame frames: it doesn't paint on top, it doesn't lay a veil, it doesn't cover exactly what you have to read.
A subscription or a payment
Speechify charges monthly because it has to charge monthly: every time you hit play, there is a rented GPU on the other side costing money by the second. That's arithmetic, not greed.
Filmina doesn't have that arithmetic because it doesn't have that other side. The model runs on your chip and the electricity is on you, like exporting a video. That is why it is paid for once: US$ 39 at launch, or US$ 19 with a student email, and the copy you installed keeps opening PDFs even if we stop existing tomorrow.
In numbers: a year of Speechify Premium is around US$ 139, and the following year another US$ 139. Filmina is one payment and that's it.
Where your notes are while you listen to them
To read you a PDF in the cloud, it has to be uploaded to the cloud. With a newspaper article it makes no difference. With course material, a thesis draft or a report from work, it might.
In Filmina there is nothing to upload: the file stays in your folder, the model reads it from there and the audio is saved beside it. It works on a plane, in the countryside and with the campus wifi down.
What you lose if you choose Filmina
This is a comparison, not a brochure. What Speechify has and Filmina doesn't:
- macOS only, version 13 or later. There is no iPhone, no Android, no Windows, no browser extension.
- It only explains in English and Spanish.
- PDFs only. It doesn't read web pages, emails or photos of a handout.
- The first pass takes time. Writing and synthesising a slide takes a few seconds; the one you are looking at is generated first and the rest is prepared behind it, but it isn't instant like hitting play on text that is already there.
- A scanned PDF is no use if it has no text layer. Speechify, with its scanner, does solve that.
- It doesn't interpret charts. It tells you there is a figure and tells you to look at it.
Which one suits you
Speechify, if your thing is dead time
You listen to long texts while travelling, you switch devices a lot, you read in several languages or you need it to read anything, not just class notes.
Filmina, if your thing is understanding
You have the slide in front of you, the problem is not reading it but that it tells you nothing, and you'd rather pay once and keep your files on your own computer.
They aren't mutually exclusive either. Some people listen to the whole chapter with a reader on the way to class and then sit down with Filmina to go through the slides they didn't understand.
Other alternatives to Speechify
If you are comparing readers, these are the other names that will come up:
- ElevenReader, by ElevenLabs. The most realistic voices on the market and a generous free plan. Same approach as Speechify: read, not explain.
- NaturalReader. The classic, widely used in accessibility, with a free web version.
- Adobe Acrobat includes «Read Out Loud» for free, with the system voice. It sounds worse, but you already have it installed.
- NotebookLM and ChatGPT are playing a different game: they don't read, they summarise. We compare them separately.
Frequently asked questions
Is Filmina an alternative to Speechify?
Only if what you wanted from Speechify was to understand a class PDF. If what you wanted was to listen to texts on the bus, on your phone and in several languages, Filmina doesn't replace that: it is Mac only and PDFs only.
Does Speechify have a free version?
Yes. The free tier includes around ten more robotic synthetic voices and reads up to 1.5x. The good voices, 5x speed and the AI extras are on the paid plan.
Does Filmina work without internet?
Yes, completely. The model that writes the explanation and the one that provides the voice are installed on your Mac, so there is nothing to ask for outside.
Does Filmina read the PDF out loud?
Not in those words. What you hear is an explanation of the slide, not the slide. If what you want is the literal reading, a classic reader suits you better.
How much does Filmina cost?
US$ 39 launch price until October 21, and US$ 19 with a student email. After that they go to US$ 78 and US$ 38. It is a one-time payment, no subscription, and you can try it free with the first ten documents, up to slide 100 of each one. See the pricing.
Keep comparing
- Filmina vs NotebookLMGoogle's two-host podcast, against the explained slide.
- Filmina vs ChatGPTWhat happens when you upload a sixty-slide PDF to the chat.
- Why I made FilminaWhere it came from: an exam, a hundred slides and no tool that did the job.
by @ignaprados