tulpy

Comprehensible Input (Krashen's Input Hypothesis) means language acquisition happens when you understand messages slightly above your current level (i+1). Research shows that comprehensible and compelling input—content you largely understand with a bit of novelty—drives acquisition better than drills or explicit instruction. Tulpy automatically adjusts difficulty so you're always in this "sweet spot." (Krashen, 1985; Krashen, 2005)

Spaced Repetition means reviewing material at increasing intervals—right before you forget—to strengthen recall over time. This "spacing effect" is one of the most robust findings in cognitive psychology: distributing practice leads to dramatically better long-term retention than cramming. Tulpy uses algorithms to optimize when you review both vocabulary and grammar patterns, scheduling them at the moment that strengthens memory most. (Cepeda et al., 2006; Tabibian et al., 2019)

According to linguistics research, grammar is learned passively through exposure to language in use, not through explicit rule memorization—your brain extrapolates patterns over time and intuits constructions. When you see patterns repeatedly in meaningful contexts, you internalize them automatically—this is how native speakers acquire grammar and how you become fluent. Tulpy focuses on exposing you to patterns in real texts, so you internalize grammar naturally without studying rules. (Ellis, 2006; Tomasello, 2003)

Contextual learning shows you how words are actually used—what they pair with and what they mean in real situations. Research consistently shows that vocabulary learned through repeated encounters in varied contexts leads to better retention, deeper understanding, and successful transfer to real communication compared to memorizing isolated word lists. Tulpy shows you words inside real sentences across different topics, so you learn how they're actually used. (Nation, 2013)

Learning from content that is interesting to the learner sustains attention and motivation. Studies show that learners who choose compelling texts read more and acquire more language. Interest triggers deeper processing and better encoding in memory. Krashen calls this "compelling input"—when content is so interesting you forget you're learning a language. Tulpy learns your interests from your taps and likes to deliver content you care about. (Krashen, 2005; Deci & Ryan, 2000)

Tulpy starts at a comfortable level and uses signals from your behavior—how often you tap words, request sentence translations, and which cards you like—to intuit your ideal level over time. It automatically refines the difficulty, keeping you in the optimal learning zone (mostly familiar + a bit new). You don't need to configure anything—Tulpy learns from your interactions and adapts in real time.

Currently: English, Russian, Ukrainian, Dutch, Spanish, French, German, Italian, Portuguese, Serbian, Chinese, Japanese, and Korean. We have plans to expand coverage to other highly requested languages in the future.

Tulpy is open now — you can start learning right now. We're still in beta and rolling out to more people over the coming months.