Portuguese flashcards: answer the Brazilian question first

You learn pequeno almoço from a deck, fly to São Paulo, and order breakfast with a phrase nobody there uses. They say café da manhã.

Nothing was wrong with the card. It was a European Portuguese card, and you were in Brazil, and no page in your search history ever asked which one you wanted.

That's the shape of this whole query. Flashcardo's Portuguese page is European Portuguese, stated once, in the opening line, then never carried into any of the 16 topic sets or the 1,000-word frequency list beneath it. PortuguesePod101 offers 1,500 illustrated words across 60 topics and never names a variety anywhere on the page. Both read as if Portuguese were one thing.

The fork you have to take first

The vocabulary overlap between the two varieties is large. It's the sound that splits them, and the split runs in the direction that hurts a flashcard learner most.

European Portuguese reduces unstressed vowels until they're barely there, and in ordinary speech they drop out completely. Telefone has four syllables on the page and arrives in your ear closer to "tlfón", with the first two vowels gone and the last one shrunk to almost nothing. That produces consonant clusters that appear nowhere in the spelling.

So a learner who built 400 cards from Brazilian audio and then lands in Lisbon can read every word on the menu and catch maybe half of what the waiter said. The words were fine. The ear was trained on a different signal.

Brazil is where the material is, which is worth saying plainly if you're undecided and have no particular reason to choose. More speakers, more television, more tutors, more decks. Brainscape's Brazilian A1 course exists because that's the larger market.

Pick by where you're going and who you'll be talking to, then make the deck name the variety before you spend anything. If a page won't tell you, that's information about the page.

There's a second reason this matters, and it's about the FSI numbers people quote at each other. The State Department puts Portuguese in Category I at 24 weeks, faster than Spanish and French at 30. Fine. But the six-week gap says nothing about listening, and European Portuguese is where most learners report the wheels coming off early.

What each one costs, and what you get

Not one page in this search prints a competitor's price. It costs me nothing to do it, so here is everything I could confirm on each company's own page, read 21 September 2026.

Three things in that table deserve more than a cell.

Brainscape is the one to beat here, and I'd rather say so. It publishes all 14 deck names with per-deck card counts before you pay, it names a price in plain words, and it runs separate Brazilian and European A1 courses, 497 and 501 cards, at the same $8. Two of its Brazilian decks are Pronunciation Gotchas and Accentuation & Emphasis, which is exactly the thing that breaks Portuguese beginners. Nobody else on this search built for that.

Its numbers are where it loses me. The same page says "Learn six months of Portuguese in two weeks!", "learn TWICE as efficiently" and "scientifically proven to help you learn faster", with no study, no author byline and no date attached to any of them. Those sentences sit next to the $8 ask. The one checkable figure on the page, "about 60 to 100 hours of study" to reach A1, is also uncited.

PortuguesePod101 hides the product and the price. You can't see a single card, topic name or deck before creating an account, "FREE" is the loudest word on the page, and the business behind it is a paid subscription the flashcard page never mentions. It also sorts you into visual, tactile and auditory learners, which is a framing the evidence doesn't support, and then never confirms the cards have audio at all.

Why Portuguese punishes a silent card

Here's the argument for spending money on this, and it isn't about card counts.

Seymour, Aro and Erskine tested foundation reading across 13 European orthographies in 2003. Their finding, verbatim: children in most European countries "become accurate and fluent in foundation level reading before the end of the first school year. There are some exceptions, notably in French, Portuguese, Danish, and, particularly, in English."

That's a study of children reading their first language, so carry the property forward rather than the result. The property is that Portuguese spelling maps onto sound less predictably than Italian or Spanish does, and that's true whoever is learning it.

Now add Bassetti and Atkinson (2015), who found the written form still bends pronunciation in learners with years of instruction behind them. A word you've only ever read in Portuguese is a word you'll mispronounce and, worse, fail to recognize when somebody says it at speed.

So audio on a Portuguese card is load-bearing in a way it isn't for Spanish. PortuguesePod101 sells to "auditory learners" and never confirms its flashcards have sound. Flashcardo's printables are paper. That leaves Brainscape, which states audio on every expression, and it's a real point in its favour.

One honest note on ours, since I'm about to describe it: Verbamor's card audio is synthesised speech generated per card with a native-accent instruction, not a human recording. Babbel voices its content with native speakers and on that specific axis Babbel wins.

What goes on one card

Every page on this search describes its cards. Here's one, so you can argue with the actual thing.

Two design choices there have sources behind them, and one has a caveat the marketing pages skip.

The blank. Producing a word beats recognizing it in a lineup. Karpicke and Roediger taught students 40 Swahili-English pairs and tested a week later: the conditions that kept retrieving a pair recalled about 80% of the list, and the conditions that only restudied recalled 36% and 33%. Students in all four conditions predicted they'd recall about 50%, and the predictions didn't differ. They couldn't feel which method was working.

The image, with a caveat. PortuguesePod101's whole pitch is that pictures beat word lists. Carpenter and Olson tested precisely that on foreign vocabulary in 2012 and found pictures did not initially beat native-language translations, and made learners more overconfident about words they'd seen with a picture. The benefit appeared only once that overconfidence was corrected by retrieval practice or an explicit warning. An image helps when it isn't allowed to make you feel like you already know the word.

The week-four arithmetic

All three competitor pages stop at acquisition. Here's the part that decides whether you're still doing this in November.

Your review load comes from your new-card rate, not the catalog size. Anki's manual puts it plainly: "If you are consistently learning 20 new cards a day, you can expect your daily reviews to be roughly about 200 cards/day." The same manual warns that more than one user "excitedly studied hundreds of new cards over their first few days" and then drowned. At 5 new cards a day the arithmetic survives contact with a job. At 20 it's a second one.

Which means 1,500 illustrated words was never a target you were supposed to finish, and reading it as one is how people quit in week three.

Grade honestly, and fail cards freely. An honest "again" costs one extra review tomorrow. A generous "good" on a word you only recognized after the flip costs 3 weeks of silent decay plus a scheduler whose model of that word is now wrong. The full rubric is here, and the bright line is whether the word came out of your head before you looked.

Run input alongside it. Nation (2006) found unassisted comprehension of spoken text at 98% coverage takes 6,000 to 7,000 word families, and he describes what 95% coverage feels like in a way I think about a lot: roughly 7 unknown words per minute of speech at a normal 150 words a minute. No deck closes that alone, and with European Portuguese vowel reduction on top, listening hours are the thing you can't substitute. The split is argued out in comprehensible input vs. flashcards.

Who should not pay for any of this

Three groups, and I mean these literally rather than as a rhetorical move.

Flying to Lisbon in two weeks? Print Flashcardo's PDFs. They're free, they come in both directions and in duplex, and a fixed list of survival words on paper is the correct tool for a trip with a date on it. Spaced repetition aims at remembering in six months, which is the wrong target for next Thursday. Just check that European Portuguese is the variety you want, because that's what you'll get.

Enjoy making cards? Use Anki. Free on Windows, macOS, Linux and Android, US$24.99 once on iOS, no subscription. It ships no Portuguese of its own, which is the catch and the point: you buy an engine and supply the words. Turn FSRS on at the bottom of the deck options page, because Anki still defaults to SM-2 from 1987 and plenty of articles wrongly say otherwise. For someone who likes the craft, Anki is the right answer to this entire page.

Working through a structured A1 syllabus? Brainscape at $8 beats us on price and on scope. 497 cards ordered by CEFR level, with pronunciation decks, is a sensible thing to buy if you want somebody else's curriculum. We don't sell one.

So what's the case for paying us more than double Brainscape's price? It starts at the point where you're having real conversations in Portuguese, losing words out of them, and finding that no catalog contains the specific words you keep losing. A static deck of 497 or 1,500 cards teaches somebody else's vocabulary. Verbamor records or reads what came up in your own lesson and turns the words and phrases you actually reached for into cloze cards with audio and an image, scheduled by FSRS. Gabriel Wyner, who has made more language cards by hand than almost anyone, puts card creation at 2 to 3 hours a week to sustain a 25-a-day habit. That labour is the line item you're buying out.

The starter deck is 670 Portuguese cards, ordered so that members of a confusable set stay apart: across the shared word list there are 199 tagged sets, the largest being 33 numbers, 24 body parts and 11 colours. Words from one category are harder to learn when they arrive together, which Tinkham found in 1993 and 1997, Waring replicated in 1997, and Nation turned into a guideline in 2000. Spacing shrinks that penalty (Nakata and Suzuki, 2019) without removing the reason to spread the set out.

Honest limits. It's iOS only, no Android and no web. It teaches you nothing new, it only holds what your lesson taught you, so it won't carry an absolute beginner through their first thousand words. It's Brazilian Portuguese. And it's 5 languages rather than a general flashcard tool.

Questions people actually ask

---

Sources

Every load-bearing claim Verbamor makes is traced to its paper on the research page.