How memorization works
A 3-step flow built on decades of cognitive-science research. See it, recall it, test it — then let spaced repetition keep it there.
Your passage is broken into bite-sized chunks. For each chunk you see it, hide it, recall it from memory, then check. Cumulative recall every few units locks the order in — not just the words.
Three drill modes harden the memory: full runs from the beginning, random starts that drop you into the middle of the passage so you can't lean on sequence alone, and weak-spot focus on the chunks you keep missing. If you stall, progressive hints reveal two words at a time.
Type the whole passage back from memory. As you type, each word turns green when it's correct and red the moment it's off, so you get live feedback without waiting until the end. We compare your final answer to the original word-for-word — no AI paraphrasing, no fuzzy matching — and highlight exactly which words you missed so you can run it again.
And then it sticks
Without review, memory decays along the forgetting curve described by Ebbinghaus1 — most of what you learn slips within days. Spacing reviews at expanding intervals flattens that curve. After each review, Memgrain schedules the next one a little further out (or sooner if you struggled), using an SM-2 style algorithm5.
The science
Pulling information out of memory strengthens it far more than re-reading. The testing effect, repeatedly confirmed in controlled studies, is the engine of every step in this flow.
Working memory holds only a handful of items at once. Breaking a passage into small chunks respects that limit so each unit can actually be encoded before moving on.
Reviews spread out over time produce dramatically better long-term retention than the same reviews crammed together. Memgrain spaces every saved text automatically.