A real curriculum.
Not random practice.
196 grammar lessons in a prerequisite graph, 866 diagnostic probes, 920 vocabulary items, 350 verbs across 12 tenses — all calibrated to CEFR standards and paced for adults.
Grammar, vocabulary, conjugation, conversation, reading, listening, review, and immersion content — everything a self-study learner needs, synchronised into one daily loop. No need to piece together five different apps.
Nine English Advantage lessons are woven through the curriculum — teaching English speakers to leverage cognates, structural parallels, and false-friend awareness at exactly the right moment in their grammar journey. See what you already know →
Everything in one place.
Grammar, vocabulary, verb drills, conversation, reading, listening, review, and curated immersion content — eight activities that read your level, your vocabulary theme, and your recent mistakes. No need to juggle five different apps.
Core loop — required daily
Learn
New vocabulary + diagnostic sweep + spaced drill. 3-8 items/day, ranked by urgency.
Lesson
196 grammar lessons with AI-generated content. Prerequisites enforced — no skipping ahead.
Conjugation
350 verbs across 12 tenses. Micro-lesson + paradigm drill + real sentence. Gated by lesson progress.
Speak
AI-powered conversation with 8 recurring personas. Inline correction on every turn. Scenarios aligned to your current vocabulary theme.
Daily input and maintenance
Review
FSRS-6 spaced repetition across 5 modalities: recognition, cued, written, spoken, spontaneous.
Reading
Daily micro-texts at your level — WhatsApp chats, notices, emails, narratives. Comprehension questions calibrated to your grammar scope.
Listening
Dictation from your own vocabulary sentences + listen-and-answer comprehension. Built from what you are actually learning.
Discover & Shadowing
Daily vocal shadowing studio to build motor speech memory, alongside living culture, podcasts, and real media from Spain, matched to your level and current lesson.
From zero to fluent. Every week planned.
A1 — Beginner
14 weeksGender, articles, ser/estar, present tense, gustar, questions, negation, ir a + infinitive
By week 14 you can order a coffee, introduce yourself, ask directions, and survive basic daily situations in Spain.
A2 — Elementary
12 weeksPast tenses (preterite, imperfect, present perfect), object pronouns, reflexive verbs, imperative, future, conditional
By week 26 you can tell stories about the past, make plans, give opinions, and handle most everyday situations.
B1 — Intermediate
21 weeksSubjunctive (7-lesson chain), compound tenses, passive voice, reported speech, discourse markers, repair strategies
By week 47 you can argue a point, tell a structured story, navigate formal situations, and self-correct mid-sentence.
B2 — Upper-Intermediate
15 weeksImperfect subjunctive, unreal conditionals, sequence of tenses, formal/colloquial register, collocations
By week 62 you can debate, write formally, use irony, and handle any register from gestoria clerk to bar slang.
C1 — Advanced
12 weeksFine subjunctive shades, hedging, nominalisation, group conversation, self-correction, discourse cohesion
By week 74 you can hold your own in a group debate, write academic text, and adapt your register on the fly.
Every part of Suelto has a paper behind it.
We didn't invent these principles. We read them, stress-tested them on one learner first, and built our own implementation. This is the paper trail: what each part of the product rests on, and where you can read the original finding.
Avoidance is invisible in error counts
In 1974, Schachter showed that learners who avoid a structure make fewer errors and therefore look more proficient than they are. Grading what learners produce misses what they quietly route around. The sweeps, hesitation timing and avoidance detection exist to measure exactly that blind spot.
Words are learned in company
The lexical approach and formulaic-language research converge on one point: fluency runs on prefabricated chunks, not isolated words. Suelto teaches 920 collocations and formulaic sequences, and involvement-load research shapes how each new item is introduced and retrieved.
Mastery gates, not calendar gates
Bloom's mastery learning holds that achievement should gate progress, not time on the calendar. The 196-lesson prerequisite graph enforces exactly that, level by level against the CEFR descriptors.
Automaticity has a speed threshold
Skill-acquisition theory separates knowing a rule from retrieving it under time pressure: ACT theory models the shift from declarative to procedural memory, and L2 research operationalises it as fluency under pressure. A correct form at four seconds is not yet automatic, so Suelto times every paradigm form and gates mastery on speed.
Input alone is not enough; output must be pushed
Swain's output hypothesis: producing language forces noticing, tests hypotheses and drives syntactic processing that listening alone never demands. Long's interaction hypothesis adds negotiation of meaning. Conversations in Suelto are production tasks by design: you respond before being told what to say, and the counterpart negotiates at your level.
Prompts beat recasts
Two 2010 meta-analyses converged: corrective feedback produces durable gains, and prompt-style feedback that pushes the learner to self-repair outperforms plain recasts. Corrections across 15 categories follow that shape: metalinguistic, in context, feeding tomorrow's practice instead of ending the turn.
Forgetting is the schedule, retrieval is the event
The testing effect is among the most replicated findings in learning science: retrieving strengthens memory more than re-exposure. Distributed-practice and desirable-difficulties research say when. Every Suelto review is an active retrieval event, spaced and interleaved.
A three-variable model of each memory
Modern schedulers descend from Wozniak's two-component model, which split memory into stability and retrievability and was later extended with difficulty into the DSR family. The FSRS scheduler we run is an open-source DSR implementation fit to each learner's review history; Duolingo's published half-life regression showed the same architecture works at scale.
Calibrated input, plus shadowing for the ear
Krashen's input hypothesis: acquisition needs messages one step beyond the current level. Every passage, transcript and conversation reply is generated inside your grammar gates, at i+1. And the Shadowing Studio follows the evidence that speech-to-speech repetition trains bottom-up phonological processing.
Continue the text; the alignment does the teaching
A substantial literature has grown around the continuation task: learners read an unfinished story and write its ending, and aligning with a stronger text drives uptake of words and structures. Its theoretical base is Pickering and Garrod's interactive-alignment account of dialogue. The Lab is a continuation-task engine: our scene, your continuation, corrections in context.
Fluency is measurable, and read-aloud flatters it
L2 fluency research operationalises speech as rate, pause ratio and mean length of run, and shows these respond to repetition practice. Suelto tracks exactly those metrics from your free speech, and excludes read-aloud, which measures decoding rather than production.
External rewards quietly tax internal ones
A meta-analysis of 128 experiments found that tangible, expected rewards reliably undermine intrinsic motivation. Streaks and XP are engineered obligations; adults who already want the language do not need them. We removed the mechanism, not the motivation.
Communicative competence is more than grammar
Canale and Swain's model split communicative competence into grammatical, sociolinguistic and strategic components; the CEFR later operationalised functional competence as can-do descriptors. The 506 function checks test the acts nobody grades: refusing politely, hedging, repairing.
How to read this list
A citation is credit and a way for you to verify; it is not an endorsement of Suelto by its authors. We read this research and built independent implementations: no text, exercise, table or curriculum is reproduced from any source, and the engineering decisions, the diagnostic engine, the nightly recompute, the personas, are ours. Where a link is missing, the finding predates stable publisher URLs; the citation stands.
Selected sources
- Anderson, J. R. (1996). ACT: A simple theory of complex cognition. American Psychologist, 51(4), 355–365.
- Bjork, R. A., & Bjork, E. L. (2011). Making things hard on yourself, but in a good way: Creating desirable difficulties to enhance learning. In Psychology and the Real World (pp. 56–62). Worth.
- Bloom, B. S. (1968). Learning for mastery. Evaluation and Comment, 1(2), 1–12.
- Canale, M., & Swain, M. (1980). Theoretical bases of communicative approaches to second language teaching and testing. Applied Linguistics, 1(1), 1–47.
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354–380.
- Council of Europe (2001). Common European Framework of Reference for Languages. CUP; Companion Volume (2020).
- de Jong, N., & Perfetti, C. A. (2011). Fluency training in the ESL classroom. Language Learning, 61(3), 877–914.
- Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125(6), 627–668.
- DeKeyser, R. (Ed.) (2007). Practice in a Second Language. Cambridge University Press.
- Ebbinghaus, H. (1885). Über das Gedächtnis. Duncker & Humblot.
- Hamada, Y. (2016). Shadowing: Who benefits and how? Language Teaching Research, 20(1), 35–52.
- Hu, M., & Nation, P. (2000). Unknown vocabulary density and reading comprehension. Reading in a Foreign Language, 13(1), 403–430.
- Kang, E., & Han, Z. (2015). The efficacy of written corrective feedback in improving L2 written accuracy: A meta-analysis. The Modern Language Journal, 99(1), 1–38.
- Karpicke, J. D., & Blunt, J. R. (2011). Retrieval practice produces more learning than elaborative studying with concept mapping. Science, 331(6018), 772–775.
- Krashen, S. (1982). Principles and Practice in Second Language Acquisition. Pergamon; The Input Hypothesis (1985). Longman.
- Laufer, B., & Hulstijn, J. (2001). Incidental vocabulary acquisition in a second language. Applied Linguistics, 22(1), 1–26.
- Lewis, M. (1993). The Lexical Approach. Language Teaching Publications.
- Li, S. (2010). The effectiveness of corrective feedback in SLA: A meta-analysis. Language Learning, 60(2), 309–365.
- Long, M. H. (1996). The role of the linguistic environment in second language acquisition. In Handbook of Second Language Acquisition (pp. 413–468). Academic Press.
- Lyster, R., & Ranta, L. (1997). Corrective feedback and learner uptake. Studies in Second Language Acquisition, 19(1), 37–66.
- Lyster, R., & Saito, K. (2010). Oral feedback in classroom SLA: A meta-analysis. Studies in Second Language Acquisition, 32(2), 265–302.
- Pickering, M. J., & Garrod, S. (2004). Toward a mechanistic psychology of dialogue. Behavioral and Brain Sciences, 27(2), 169–190.
- Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning. Psychological Science, 17(3), 249–255.
- Rohrer, D., & Taylor, K. (2007). The shuffling of mathematics problems improves learning. Instructional Science, 35(6), 483–503.
- Schachter, J. (1974). An error in error analysis. Language Learning, 24(2), 205–214.
- Segalowitz, N. (2010). Cognitive Bases of Second Language Fluency. Routledge.
- Settles, B., & Meeder, B. (2016). A trainable spaced repetition model for language learning. Proceedings of ACL 2016.
- Swain, M. (1985). Communicative competence: Some roles of comprehensible input and comprehensible output. In Input in Second Language Acquisition (pp. 235–253). Newbury House; Three functions of output in second language learning (1995). OUP.
- Wang, C., & Wang, M. (2015). Effect of alignment on L2 written production. Applied Linguistics, 36(5), 503–526.
- Wozniak, P. A., Gorzelanczyk, E. J., & Murakowski, J. A. (1995). Two components of long-term memory. Acta Neurobiologiae Experimentalis, 55.
- Wray, A. (2002). Formulaic Language and the Lexicon. Cambridge University Press.
You never see what you haven't studied.
23 grammar gates monitor every piece of content the system generates. If you haven't passed the imperative lesson, no AI-generated text will use imperative forms — not in reading, not in listening, not in conversation. This is enforced across 16 content generation points in real time.
Siempre hago caso a mis padres.
present tenseHaz caso a tu madre cuando te lo diga.
imperative + subjunctiveSame vocabulary item. Different grammar. Calibrated to you.
grammar lessons
diagnostic probes
verbs drilled
tenses
months A1 → C1
Start with a free diagnostic.
12 prompts. About 10 minutes. See exactly where your Spanish is.