CEFR coverage A1 → C2 Full Council of Europe framework alignment to mastery level. | All 6 levels with sub-bands | ~A1 → ~B1 | ~A1 → B1 | ~A1 → B1 | ~A1 → B1 | ~A1 → B1 | | | | — | — |
Per-skill ELO proficiency (8 skills) Reading, writing, listening, speaking, vocab, grammar, pronunciation, translation rated independently with confidence bands. | | —Single XP level | — | — | — | — | — | — | ~Vocab-only level | — | — |
24-factor adaptive insights with NL commentary Engagement, velocity, retention, balance, quality, application — every factor with plain-language explanation. | | ~Streak + XP only | ~Basic stats | — | — | — | — | — | ~Words-known stat | — | — |
AI conversation tutor (free-form, persona-driven) Real back-and-forth in target language with implicit error recasting and structured coaching after. | | ~Roleplay add-on, scripted | — | — | — | — | | — | — | — | — |
Adaptive tutor difficulty mid-conversation Reads engagement signals (struggle markers, response length, vocab breadth) and modulates complexity in real time. | | — | — | — | — | — | | — | — | — | — |
AI free-response grader with rubric Free writing + speech transcripts graded against a 5-criterion rubric with diff annotation. | | ~Limited | — | — | — | — | | | — | — | — |
Browser TTS at multiple speeds for every exercise Half-speed, native, and stretch playback wired to dictation, shadowing, read-aloud. | | ~Native only, single speed | ~Native only | | | | | | | | |
Phoneme-level pronunciation feedback Per-phoneme red/green diff with articulatory hints (e.g., 'try a tap [ɾ] not English /ɹ/'). | | ~Overall % only | ~Word-level | ~Speech analysis but opaque | — | — | | — | — | — | — |
Voice-capture personal references (compare against your past self) Record a phrase once, then every future attempt compares to your own baseline. IndexedDB-private. | | — | — | — | — | — | — | — | — | — | — |
In-browser graded readers with click-to-gloss + Add-to-deck Cervantes, Kafka, Dante, Sōseki, Lu Xun — public-domain and CEFR-tagged. | 15+ texts across 8 languages | ~Stories module, scripted | — | — | — | — | — | — | Strongest competitor here | — | — |
Bilingual sentence-aligned reading mode L1 + L2 side-by-side with hover glosses on every word. | | — | — | — | — | — | — | — | | — | — |
Writing analytics (TTR, sentence CV, connective density, cohesion suggestions) Pure-functional engine analyzes your prose and suggests cohesion fixes. | | — | — | — | — | — | | | — | — | — |
Translation workbench with per-clause linguistic analysis Multiple renderings (literal/idiomatic/formal/literary/academic), per-clause reasoning, false-friend warnings. | | — | — | — | — | — | — | — | — | ~Single literal output | ~Strongest 1:1, but no analysis |
Travel mode with curated phrase packs + live two-way conversation + 'show this' large text Built for travelers: 12 situations, 27+ phrases per language, instant TTS, two-way mic interpretation. | | ~Phrasebook in some languages | ~Travel module | — | Audio-only travel modules | — | — | — | — | Live conversation; no curation | —No conversation |
Cross-language interference detection Surfaces false friends across your studied languages (PT/ES, JA/ZH, FR/EN, DE/EN, IT/ES) before you fall in. | | — | — | — | — | — | | — | — | — | — |
Forgetting-curve forecast (predict at-risk cards next 7-30 days) Per-card stability projection so you review what's slipping, not what's stable. | | — | — | — | — | ~Has SRS but no forecast | — | — | — | — | — |
Self-assessment calibration tracker (Brier score) Are your confidence ratings honest? Metacognition is the strongest predictor of expert learners. | | — | — | — | — | — | — | — | — | — | — |
Coverage map by CEFR (mastered / learning / seen) Visual breakdown of curated content traversed at every level. | | ~Tree of completed nodes ≠ coverage | ~Lessons completed | ~Partial / weaker | — | | — | ~Partial / weaker | | — | — |
12 distinct exercise types (vs 4-5 standard) Translation, cloze, dictation, MC, word-order, read-aloud, free-speech, free-writing, comprehension, conversation, transformation, shadowing. | | ~~5 types | ~~5 | ~~6 | ~Audio drills only | ~~4 | | ~~5 | ~Reading + listening | — | — |
Mature gamification (mastery milestones, no streak guilt) 12 milestones tied to real progress. No daily-reset streaks, no 'streak lost' anxiety. | | —Streak guilt-tripping is the business model | ~Partial / weaker | ~Partial / weaker | — | ~Streaks | — | —Streaks | ~XP | — | — |
University cohort + instructor dashboard + auto-grading FERPA-aligned, anonymous leaderboards, CSV roster import, rubric override. | | ~Duolingo for Schools, light | ~B2B teams | | ~Library partnerships | ~Partial / weaker | — | ~B2B | — | — | — |
PWA offline mode (app-shell + cached content) Service worker + manifest for true offline learning on a flight or rural commute. | | ~Limited offline | | | | | — | | | | ~Partial / weaker |
27 languages including classical (Latin, Ancient Greek, Sanskrit) Modern + classical + heritage for university programs that teach them. | | | ~14 | ~25 | ~50+, audio quality | | | ~12 | | | |
Free of ads — forever No ads, no streak guilt-tripping, no manipulative push notifications. | | —Ad-supported free tier | | | | —Free tier limited | | —Free tier limited | | | |
Free for accredited universities (≤50 seats) Non-commercial institutional use is free — hand it to a class today. | | ~Free for K-12, paid for higher-ed | — | —Education-tier paid | — | — | — | — | — | | |