Mnestica starts with a concrete loop: provide source text, generate cards, save them, and complete a review. Notes, reading items, and repair tools add context around that loop without claiming to measure a learner's long-term memory.

AI-generated flashcards

In Create, you can paste text or select an eligible internal note or reading item, choose Q&A or cloze cards, and choose a card count. Mnestica sends the request through its configured AI generation path. You review the returned cards before saving them to a new or existing deck.

Saved cards use Mnestica's SM-2-inspired scheduler. Their Forgot, Hard, Good, and Easy ratings update intervals, ease factors, and due dates.

Connected notes

Notes can record a concept, example, misconception, question, analogy, or application. Notes can link to other notes and to flashcards. During study, you can capture an insight and keep it associated with the card that prompted it.

These links are stored product relationships. They do not imply that the linked material has been understood or retained.

Reading inbox

Reading items let you keep source material in a staged queue, work through blocks, and promote selected material into notes or card-generation inputs. You decide what to promote and what to defer.

On-demand card repair

Fix card always lets you edit the question and answer directly. From a study card or an open care case, review evidence can also support a suggestion to rewrite, split, suspend, or retire the card. Mnestica shows the proposed change and rationale. The card changes only after you approve the action.

Care cases are deterministic product signals based on stored practice state. They can draw attention to repeated failures or maintenance needs, but they are not a diagnosis of why a person forgot.

Memory Studio

Memory Studio reads current product data and displays:

  • practice-target inventory;
  • scheduled due burden;
  • queue composition;
  • open care cases; and
  • repair actions for cards that need attention.

Forecasted burden is a count of items currently due in future buckets, not a promise about how much work a learner will complete.

The result is one inspectable place to decide what to review, read, or repair next. It is a workflow view, not a claim that the software guarantees durable memory or educational outcomes.