Mnemio

Concept

Mnemio is a platform for deliberate memorization of anything — from interview questions to foreign vocabulary. At its core is not a timer or passive re-reading, but the FSRS-6 spaced-repetition algorithm that decides what to show and when. Content is organized into two levels — Collections and Decks — and answers are stored as rich text so a card can hold code, lists, and formatting rather than a single string.

Engineering

• Strict Feature-Sliced Design: app / pages / widgets / features / entities / shared layers with one-way imports and per-slice public APIs.
• React 19 with React Compiler — automatic memoization without manual useMemo/useCallback; type-safe routing on TanStack Router.
• Server state on TanStack Query with Suspense and skeletons, client state on Zustand; form and data validation with Valibot.
• Spaced-repetition domain on ts-fsrs (FSRS-6): stability, difficulty, card states, and an append-only revlog as the source of daily metrics.
• Supabase for database, auth, and Row-Level Security; AI answer grading via user-provided OpenRouter keys proxied through Edge Functions.
• Deployed on Cloudflare Pages; groundwork laid for personalized FSRS weight optimization in a Web Worker (WASM).