Methodology
How dailyBe’s content is made, and where it can go wrong
Where the app’s sentences and audio come from, how you report a mistake, what the app measures, the research the method builds on, what the app doesn’t do, and how the blog’s articles are put together.
On this page9
The short version
- Sentences and audioGenerated by artificial intelligence.
- MistakesYou report the error in the app, and the sentence goes through human review.
- ProgressEach word is measured as correct answers ÷ times seen. Once every word in a sentence is learned, it leaves active study.
- ReviewA ladder of intervals of 1, 1, 3, 5, 7, 14, 28 and 54 days: right, it moves up a step; wrong, it moves down one.
- LimitsNot for complete beginners, doesn’t work offline yet, and English only for now.
- BlogThe dailyBe Team, with AI assistance and source checking. Corrections within 30 days.
Sentences and audio: generated by AI
Every example sentence and every audio clip in dailyBe is generated by artificial intelligence.
There’s no hand-written sentence bank behind the dictionary. That’s what lets you look up any word, expression or phrasal verb from any field, instead of only what’s on a ready-made list: a technical term from your job and a piece of slang from a show get the same treatment, with at least five example sentences and audio.
What that means in terms of limits
AI-generated content can be wrong in ways that aren’t always obvious. In practice:
- a sentence can sound unnatural even when it’s grammatically correct;
- the meaning in an example can be right in general but not the one you ran into in your context;
- the register can be off: an informal expression presented as neutral, or the other way around;
- the translation may not match the sentence exactly;
- an audio clip can have odd pronunciation, intonation or rhythm.
That’s why it’s worth reading more than one example before you add a word to your list. And if something looks wrong, report it: that’s how a mistake reaches a person.
Error reporting and human review
Inside the app, you can report an error in a sentence and say whether the problem is the pronunciation or the translation. Reported sentences go through human review.
Report a sentence whenever it:
- has a translation that’s wrong or doesn’t match the meaning of the word you looked up;
- has audio with the wrong pronunciation, or audio that doesn’t match the text;
- is wrong or doesn’t make sense.
Reporting is the bridge between what the AI produces and a human set of eyes: every reported sentence goes through human review. The more specific the problem, the easier it is to fix.
For now, the app doesn’t notify you when the review of a sentence you reported is done.
What the app measures: each word
dailyBe measures how you do on each word: how many times it has come up in your practice and how many times you got it right.
Here’s how it works:
- Every time a word comes up in practice, the app records whether you got it right or wrong.
- The math is simple: correct answers divided by times seen. Once that ratio passes a set percentage, the word is marked learned and stops being a focus.
- Once every word in a sentence is learned, the sentence leaves active study and the app moves the next ones on your list up.
That’s why we measure word by word, not sentence by sentence. In The book is on the table, you may have already learned the and is while you’re still learning book and table. The sentence stays in practice until the rest sticks.
What about word frequency?
The app doesn’t show any frequency score or scale. But the research on word frequency is a useful guide when you’re deciding what to look up and add to your list. A small number of words make up most of what people say, and what intermediate learners are missing tends to sit in the middle: common enough that native speakers use it without thinking, rare enough to slip through the cracks of most courses.
- Nation, 2006In the script of Shrek, the 1,000 most frequent word families account for over 81% of the words spoken; families ranked 5,001 to 6,000 account for just 0.33%, and most of them appear only once. Knowing 95% of the words in speech still leaves about 7 unknown words per minute.
- Webb & Rodgers, 2009; Ha, 2022In movies and TV, about 3,000 word families cover roughly 95% of what’s said. Getting to 98% takes somewhere between 5,000 and 7,000, depending on the study.
- Schmitt & Schmitt, 2014The band between roughly 3,000 and 9,000 word families, mid-frequency vocabulary, is needed to use the language proficiently, and teaching it is a challenge of its own.
Spaced repetition and retrieval practice
The sentences on your list move up and down a ladder of intervals of 1, 1, 3, 5, 7, 14, 28 and 54 days. Get it right and the sentence moves up a step and comes back later; get it wrong and it moves down one and comes back sooner. And the review happens in the practice modes, not as a reread.
How it works in the app
A sentence’s schedule, counted from your first practice:
Get it right every time and the sentence comes back in:
- Today1st practice
- 1 dayday 1
- 3 daysday 4
- 5 daysday 9
- 7 daysday 16
- 14 daysday 30
- 28 daysday 58
- 54 daysday 112
Get it wrong and the sentence moves down a step, back to the previous interval. Miss the review that came after 7 days, and it comes back in 5 instead of 14. Miss it right at the start, and it comes back the next day.
There’s no daily cap on reviews.
Since each word shows up in several sentences, it also comes back at different times and in different contexts, which spreads out your contact with it even more.
What the research says
Two ideas with a lot of research behind them: spread reviews out over time instead of cramming them together, and review by trying to remember instead of rereading. Here’s what the key studies found:
- Ebbinghaus, 188538 repetitions spread over three days had the same effect as 68 repetitions done in a single day.
- Cepeda et al., 2006A review of 839 assessments across 317 experiments on distributed practice: the best gap between study sessions grows with how long you need to remember.
- Kim & Webb, 2022A meta-analysis of 48 experiments with 3,411 second-language learners found a medium-to-large effect for spacing.
- Dunlosky et al., 2013Of ten common study techniques, only distributed practice and practice testing were rated high utility. Rereading and highlighting were rated low.
- Karpicke & Roediger, 2008A week later, students who kept being tested recalled about 80% of the Swahili words; those who stopped being tested recalled 36% and 33%, even in the group that kept studying the words.
To dig deeper: Spaced repetition explained and Why you forget the English you study.
Why practice by speaking, seeing and listening
Recognizing a word when you hear it and being able to use it when you talk are different levels of knowing it.
- Laufer, 1998; Webb, 2008Studies of English-as-a-foreign-language learners consistently find receptive vocabulary larger than productive vocabulary. In one of them (Webb, 2008), with strict scoring, the gap grew as the words got less frequent.
- MacLeod et al., 2010Under some conditions, saying a word out loud makes it easier to remember than reading it silently.
Practicing the same sentence in different modes is how you work both ends: understanding what you hear and read, and producing the sentence when you need it. More in I understand English but can’t speak it.
The three modes
- Speaking: you say the sentence out loud. The app uses speech recognition to show what it heard and, if that doesn’t match the sentence, gives you a pronunciation tip spelled the way a Brazilian would read it, like “róu ariu?” for How are you? or “ai xúdav nôun bérer” for I should have known better.
- Seeing: you practice the sentence from what’s on screen, which helps lock in the spelling and the words that travel together.
- Listening: you practice the sentence by ear, at the speed people actually talk.
What dailyBe doesn’t do
Being clear about what the app doesn’t do matters as much as listing what it does.
- It doesn’t teach English from scratch. The app is for people past the beginner stage who already understand 60% to 70% of a sentence. The Stories help if you’re closer to the beginning, but without a base almost everything is a gap, and it makes more sense to build the basics first.
- It isn’t a grammar or conversation course. The focus is vocabulary: the words and sentences your English is missing. Grammar and conversation are best worked on in class, with a teacher or with other people.
- It doesn’t replace classes or exposure to real English. It’s a complement: classes give you structure, and dailyBe covers the vocabulary that’s specific to you.
- It doesn’t work offline yet. Downloading sessions to practice offline is on the roadmap, with no date yet.
- It’s English only, for now. German, Spanish, Italian and French are on the roadmap, with no date yet.
- It doesn’t promise fluency by any deadline. Vocabulary grows with regular study: a few sessions a day, adding up to 20 to 30 focused minutes. This isn’t a five-minute app.
- It doesn’t choose what goes on your list. The first-launch diagnostic sets a starting direction, and the app moves the next sentences up once one leaves active study, but you decide what gets added.
- It doesn’t guarantee every AI-generated sentence is right. That’s what error reporting and human review are for.
How blog articles are made
The articles are signed by the dailyBe Team and written with AI assistance, with source checking. These standards apply to all of them.
- The articles are written with AI assistance and go through source checking: every number or research finding cited comes from a source listed at the end of the article, with a link.
- We favor primary sources: research papers, reviews and meta-analyses, reference dictionaries for word meanings, and official glossaries for technical terms.
- When a study has limits (a single subject, artificial material, a small sample), the article says so.
- Every article shows its publication date and, when revised, the date it was updated.
Found a mistake?
Let us know. Each kind of mistake has its own path.
- In a sentence or audio clip in the app: report the error inside the app.
- In a blog article or on a page of this site: email contato@dailybe.app with the link and the passage.
Corrections policy
When a mistake in an article or on a page is confirmed, we fix it within 30 days of being told. Corrections to articles get a dated update note.
Sources
The references used on this page, each with the blog article where it’s discussed.
Word frequency and vocabulary size
- Nation, I.S.P. (2006). How Large a Vocabulary Is Needed for Reading and Listening? The Canadian Modern Language Review, 63(1), 59-82Discussed in: The words your English is missing: how to get off the plateau · How many words do you need to watch TV without subtitles?
- Schmitt, N. & Schmitt, D. (2014). A reassessment of frequency and vocabulary size in L2 vocabulary teaching. Language Teaching, 47(4), 484-503Discussed in: The words your English is missing: how to get off the plateau
- van Zeeland, H. & Schmitt, N. (2013). Lexical Coverage in L1 and L2 Listening Comprehension: The Same or Different from Reading Comprehension? Applied Linguistics, 34(4), 457-479Discussed in: The words your English is missing: how to get off the plateau · How many words do you need to watch TV without subtitles?
- Webb, S. & Rodgers, M. P. H. (2009). The Lexical Coverage of Movies. Applied Linguistics, 30(3), 407-427Discussed in: How many words do you need to watch TV without subtitles?
- Webb, S. & Rodgers, M. P. H. (2009). Vocabulary Demands of Television Programs. Language Learning, 59(2), 335-366Discussed in: The words your English is missing: how to get off the plateau · How many words do you need to watch TV without subtitles?
- Ha, H. T. (2022). Vocabulary Demands of Informal Spoken English Revisited: What Does It Take to Understand Movies, TV Programs, and Soap Operas? Frontiers in Psychology, 13Discussed in: How many words do you need to watch TV without subtitles?
Memory, spacing and retrieval
- Ebbinghaus, H. (1885/1913). Memory: A Contribution to Experimental Psychology, capítulo 8 (Classics in the History of Psychology)Discussed in: Why you forget the English you study (and what makes it stick) · Spaced repetition: what it is and how it works for English
- Cepeda, N. J. et al. (2006). Distributed Practice in Verbal Recall Tasks: A Review and Quantitative Synthesis. Psychological Bulletin, 132(3), 354-380Discussed in: Spaced repetition: what it is and how it works for English
- Cepeda, N. J. et al. (2008). Spacing Effects in Learning: A Temporal Ridgeline of Optimal Retention. Psychological Science, 19(11), 1095-1102Discussed in: Why you forget the English you study (and what makes it stick) · Spaced repetition: what it is and how it works for English
- Kim, S. K. & Webb, S. (2022). The Effects of Spaced Practice on Second Language Learning: A Meta-Analysis. Language Learning, 72(1), 269-319Discussed in: Spaced repetition: what it is and how it works for English
- Dunlosky, J. et al. (2013). Improving Students' Learning With Effective Learning Techniques. Psychological Science in the Public Interest, 14(1), 4-58Discussed in: Why you forget the English you study (and what makes it stick) · Spaced repetition: what it is and how it works for English
- Karpicke, J. D. & Roediger, H. L. (2008). The Critical Importance of Retrieval for Learning. Science, 319(5865), 966-968Discussed in: Why you forget the English you study (and what makes it stick)
- Zheng, G. et al. (2026). Effects of spacing on the acquisition of explicit and implicit vocabulary knowledge. Studies in Second Language Learning and TeachingDiscussed in: Spaced repetition: what it is and how it works for English
Receptive and productive vocabulary
- Laufer, B. (1998). The Development of Passive and Active Vocabulary in a Second Language: Same or Different? Applied Linguistics, 19(2), 255-271Discussed in: I understand English but can't speak it: what to do
- Webb, S. (2008). Receptive and Productive Vocabulary Sizes of L2 Learners. Studies in Second Language Acquisition, 30(1), 79-95Discussed in: I understand English but can't speak it: what to do
- MacLeod, C. M. et al. (2010). The production effect: Delineation of a phenomenon. Journal of Experimental Psychology: Learning, Memory, and Cognition, 36(3), 671-685Discussed in: I understand English but can't speak it: what to do
Page last updated
Get the app
Start training the words your English is missing today.
For iPhone on the App Store and for Android on Google Play.
