Loading

English Apps in 2026: Your First Language Decides Which Ones Help

Most "best English app" lists are written by native English speakers who have never learned English. The rankings are useless for most readers because the question is not which app is best in the abstract. The question is which app fixes the specific problems your first language creates when it bumps into English. A Spanish-speaking accountant studying for a US visa interview and a Tokyo programmer trying to pass a TOEIC threshold do not need the same tools. A single ranked app list is the wrong format for that spread.

I am Alej Pascual. I grew up speaking Spanish in Mexico City, learned English to native-level fluency, and went on to learn nine more languages. I conduct business in English daily, lived in New York for years, and built Mynago to teach languages with the learner's actual L1 in mind. Find me on LinkedIn or as @langaholic anywhere else.

The British Council estimates roughly 1.5 billion people are studying English worldwide, pulled in from Spanish-speaking Latin America, Mandarin and Cantonese-speaking East Asia, the Japanese and Korean test-prep markets, and the Arabic, Persian, and Turkish corridors of the Middle East and Central Asia. The needs across those blocs barely overlap. This post is organized by L1. Find your section, read the framing for what your first language already gave you and what it actively hurts, then look at the apps recommended for your transfer issues.

The L1 sections

Spanish speakers: scroll down to the next section. Everyone else: skim until you find your L1, then read carefully.

The framing is the same across all sections: what your L1 gave you for free, what your L1 actively breaks, the specific apps that target the breakage. The general apps that work for everyone are at the bottom in the Universal Picks section.

L1 Spanish: the cognate-trap and schwa problem

Spanish speakers have the largest false-friend problem of any L1 coming into English, and the smallest grammar problem. The real enemies: phrasal verbs (Spanish does not really have them), the present perfect versus simple past distinction, the schwa sound (Spanish vowels are clean and constant; English unstressed vowels collapse to /ə/), and the embarrassment-pregnancy class of cognate traps.

The traps you will hit in year one: actually does not mean actualmente, eventually does not mean eventualmente, embarrassed does not mean embarazada (this one matters), carpet is not carpeta, fabric is not fábrica, pretend is not pretender. None of these are decoded by general apps. You need targeted vocabulary work and exposure to natural English where these words appear in real contexts.

Build the stack around YouGlish first. Search any English word, get dozens of YouTube clips of native speakers pronouncing it in real sentences. The schwa is the most common vowel sound in English and does not exist in Spanish. Watching native speakers reduce banana to /bəˈnænə/ instead of the cleanly-pronounced ba-na-na you instinctively produce is the fastest correction loop available.

Add Grammarly for writing. Spanish speakers tend to overuse commas (Spanish punctuation rules are looser), undertranslate the present perfect, and place adverbs in Spanish positions. Grammarly catches all three in real time. The free version is enough.

Add a phrasal-verb-focused Anki deck. Spanish has no native equivalent for put up with, get over, look forward to. Memorize them as units, not as compositional puzzles. The "English Phrasal Verbs in Use" deck on AnkiWeb is the standard.

Add Mynago for daily structured practice. Biased, but the English course is built with the learner's L1 in mind, and the dialogues surface the cognate-trap and phrasal-verb gaps that wreck Spanish-speaking professionals at the B2-to-C1 transition.

Finally, a telenovela-to-English bridge. Watch shows you have already seen in Spanish (Money Heist, Narcos, Élite) dubbed in English with English subtitles. The plot scaffolding lets you spend cognitive load on language instead of story.

If you want this guide framed entirely in Spanish, see the Spanish-language version of Mynago.

A correction from a Madrid reader: Spanish speakers do not have a single transfer profile. Castilian theta-aware speakers and Latin American speakers have meaningfully different English pronunciation issues. Spaniards collapse the English "v" into a "b" more often. Latin Americans collapse "z" into "s" more often. The general point holds, the specifics need an L2-dialect-of-L1 calibration that no app currently does well. Take the cognate-trap list above as universal across Spanish dialects, and the pronunciation drills as something you should calibrate against your own variety.

L1 Mandarin or Cantonese: articles and stress timing

The structural reset is the largest of any L1 group. Mandarin has no articles, no verb conjugation, no plural marking, and a tonal system that does not map onto English stress. You are not learning new vocabulary on top of familiar grammar. You are rebuilding the entire architecture.

The biggest single problem: articles. The a/an/the system consumes years of input even from advanced Mandarin speakers, because there is nothing in your L1 to anchor it to. Rule memorization fails. Massive reading and listening, where article use is absorbed unconsciously, is the only proven path.

The second biggest: stress-timed rhythm. Mandarin is syllable-timed where every syllable gets roughly equal weight. English is stress-timed where stressed syllables anchor the rhythm and unstressed syllables compress. PHO-to-graph versus pho-TO-gra-pher versus pho-to-GRA-phic. Your ear has to learn that distinction before your mouth can produce it.

Elsa Speak for phoneme-level pronunciation. Uses speech recognition to score individual sounds and stress placement. For Mandarin speakers the highest-value training is on /θ/ and /ð/ (the two th sounds, neither of which exists in Mandarin), word stress placement, and the /n/ versus /ŋ/ distinction at the end of syllables. Elsa isolates each.

BBC Learning English for graded input. The Pronunciation in Spoken English series and 6 Minute English are designed for the kind of high-volume listening Mandarin speakers need. Free, professional, structured by level.

Mynago for situational dialogue. Articles, prepositions, and tense are absorbed best inside real scenarios (job interview, doctor's appointment, coffee-shop order). Decontextualized drills do not stick.

Anki with frequency decks. Build vocabulary aggressively through the first 5,000 words. The "English Vocabulary Profile" frequency deck is the standard. Cantonese speakers should layer a separate deck for English words that share a sound but not a meaning with Cantonese (the short / should / shoot cluster is a known trap).

Subtitled sitcoms with English-only subtitles. Friends, The Office, Brooklyn Nine-Nine. Slower delivery, predictable plot, repeated vocabulary. Avoid Succession or The Wire until solidly B2.

L1 Japanese or Korean: the same grammar reset plus a worse sound problem

Japanese and Korean speakers face the same grammar reset as Mandarin speakers (no articles, different word order) plus a sound-system problem that is, in some respects, harder. Japanese has five vowels. English has fifteen. Korean does not contrast /r/ and /l/ in initial position. English does, and the contrast carries meaning. Both languages are SOV. English is SVO and the verb-final instinct will haunt you for years.

The underrated problem: register. Japanese and Korean both encode social hierarchy directly into verb endings. English encodes it through vocabulary choice and syntax (could you versus can you, I was wondering if versus do you). The mapping is inexact and corporate English runs on tonal subtlety invisible if you do not know to listen for it.

Mynago for register-aware dialogue. The English lessons explicitly cover formal versus casual register switching, which is a genuine blind spot for Japanese and Korean learners coming out of school-system English.

Elsa Speak for /r/, /l/, and the vowel grid. Same logic as Mandarin speakers, but Japanese and Korean speakers have an additional problem: the consonant clusters at the start of English syllables (street, splash, strength) require deliberate retraining. Elsa breaks them down phoneme by phoneme.

Cambly or italki for forced output. Japanese and Korean education systems produce learners who score well on grammar tests and freeze in conversation. The fix is conversation hours, not more grammar. Cambly is cheaper and more flexible. italki is better if you want a structured teacher.

Pimsleur audio courses for stress and intonation. Dated and audio-only. Its spaced-repetition pacing of spoken phrases is unmatched for the rhythm shift Japanese and Korean speakers need. Use it during commutes.

Avoid Duolingo's English course. It does not have a Japanese-from-English or Korean-from-English course that handles your specific issues. The reverse-direction course (English-from-Japanese / English-from-Korean) is generic and slow.

L1 Arabic, Persian, or Turkish: the underserved corridor

These three languages do not pattern together linguistically but they share a real-world feature: their speakers are underserved by the English app market. Most apps assume a European or East Asian L1. I am grouping you here with separate notes per language.

Arabic speakers. The vowel system is your main hurdle. Modern Standard Arabic has three vowels with length contrast. English has fifteen-ish vowels with no length-equivalent system. The bit/beat/bet/bat cluster will collapse into one or two sounds for years unless you actively train it. English spelling will feel deranged: through, though, tough, thought share four letters and zero pronunciation logic. Arabic spelling is largely phonetic, and your brain will reject English spelling on contact.

Persian speakers. SOV to SVO retraining (Persian, like Japanese, puts the verb at the end), no article system, the same vowel-system gap as Arabic. The good news: Persian's flexibility with word order means your brain is already comfortable holding multiple possible orderings, which speeds up SVO acquisition once you commit.

Turkish speakers. I cover Turkish in detail on our Turkish-language English landing page. Headline traps: the article system (no Turkish equivalent, around 29% of Turkish learners' written errors are article errors), present perfect versus mış confusion, vowel length (beat/bit, pool/pull, bat/bet pairs collapse into one sound), and the European loanword false-friend pool (apartman is a building, kolej is a high school, patron is a boss, pasta is cake). Phrasal verbs are also underused; memorize them as units.

The apps for all three:

Elsa Speak for vowel and consonant retraining. The vowel-system gap is the biggest pronunciation barrier for Arabic and Persian speakers. Elsa's phoneme-level feedback is the most efficient drill loop available.

Mynago for structured dialogue with article and tense focus. Article acquisition cannot be rule-memorized. It has to be absorbed through repeated exposure in context. Mynago's situational dialogues give you that exposure with a manageable difficulty curve.

VOA Learning English is underrated. Slowed-down American English news, free, with transcripts. Particularly good for Arabic and Persian speakers because it pairs spelling with sound at slow speed, which helps the spelling-pronunciation mismatch settle.

Grammarly for writing. Article errors and tense errors are exactly what Grammarly catches. The free version is sufficient.

Anki with a deck built from your own writing mistakes. Whatever Grammarly flags, add to Anki. A custom error-correction loop targeting your specific transfer issues, not generic English problems.

L1 anything else

Romance (Portuguese, French, Italian, Romanian): Spanish-speaker-like advantages with different false-friend lists. Same toolkit: Mynago, YouGlish, Grammarly, an Anki phrasal-verb deck.

Germanic (German, Dutch, Swedish, Norwegian): closer to native English than any other L1 group. Traps are the present perfect versus simple past line, false friends (German Gift means poison), and subordinate-clause word order. Skip Elsa; your baseline pronunciation is already further along than most.

Slavic (Russian, Polish, Ukrainian, Czech): articles are your main grammatical hurdle, the same as for Mandarin and Turkish speakers. Aspect-versus-tense mapping is the second issue. The /θ/ and /ð/ sounds need explicit work. Elsa for th, Mynago for article exposure, aggressive reading.

South Asian (Hindi, Urdu, Bengali, Tamil, Telugu): you have a head start from school and media exposure. Remaining work is connected speech (gonna, wanna, should've), accent clarity, and prepositions (Hindi and Urdu use postpositions). Cambly or italki for output; Elsa for accent clarity.

Vietnamese, Thai, Indonesian, Malay: SVO transfers, a real advantage. Articles and verb tenses are the grammar work. Vietnamese and Thai add a tone-to-stress retraining problem. Indonesian and Malay speakers have the lightest cognitive load of any non-European L1.

Your biggest single advantage over the Mandarin-Japanese-Arabic-Turkish bloc is that you do not need a structural reset. Push input volume.

Universal picks (work regardless of L1)

Two tools work for everyone and you should use both.

Mynago. Built-in L1 awareness, situational dialogues that match the contexts where adults actually use English, daily structure that compounds. Honest disclosure: I built it. Honest assessment: nothing else on the market is structured around the L1 you actually have.

Anki. The interface is ugly and the learning curve is real. Twenty years later it remains the most efficient vocabulary-acquisition tool ever built. Use a frequency deck for the first 5,000 words, then switch to mining your own vocabulary from real reading and listening. Quizlet and Memrise are friendlier alternatives if Anki defeats you, but they are weaker.

These two are the spine. Everything else (Elsa, BBC, Grammarly, YouGlish, Cambly, italki) is a specialized add-on for a specific weakness.

Apps to skip regardless of L1

Rosetta Stone. Expensive, the no-translation-ever philosophy frustrates adult learners, and the immersion model is weaker than just watching subtitled TV for free.

Duolingo's English course past A2. Fine for habit-building in the first three months. Past that, the gamification becomes the goal instead of the learning, speech recognition accepts almost any noise as correct, and sentences are decontextualized in a way that does not transfer to real conversation.

Generic "AI English tutor" apps that launched in the last 18 months. Most are wrappers over a single language model with no L1-awareness, no curriculum, and no spaced repetition. They feel impressive in the first session and useless by week three. The narrow exception: if you are above B1 and your bottleneck is output anxiety, an AI tutor is genuinely useful as a low-stakes daily speaking partner. Twenty minutes a day of "describe what you did today" with a system that interrupts you for grammar errors works. Below B1 the AI tutor is a distraction.

Premium IELTS / TOEFL "shortcut" courses sold via paid social ads. Exam prep is a real category but the legitimate options (British Council official prep, ETS official prep, Magoosh, E2 Test Prep) are well-known and reasonably priced. Anything sold via a paid Instagram ad with a one-week countdown timer is not legitimate. TOEFL prep apps consolidated faster than I expected. Magoosh remains strong. E2 Test Prep slipped after a UI overhaul. The "AI TOEFL coach" subcategory is mostly noise except for the official ETS prep, which you should use last week before the exam regardless of what other prep you did.

The bigger pattern

After eleven languages and decades of watching other people learn English, the most useful thing I can tell you is that the right tool depends almost entirely on what your first language already gave you and where it is now in your way. A Spanish speaker and a Japanese speaker are not in the same fight. They should not be reading the same app rankings.

Find your section above. Pick two or three apps that target your specific transfer issues. Add Mynago and Anki as the spine. Then put in the hours, because no app, mine included, replaces the hours.

The apps you start with at A1 should not be the apps you finish with at B2. Plan the handoffs.

If you want a free starting point, take the English level assessment before you spend money on anything.

Learning another language alongside English? These guides each take a different angle: