MB Logo
Home Work Projects Notes Practice About

HSK Nest

A self-hostable spaced-repetition trainer for Mandarin — all of HSK 1–9 pre-loaded, scheduled by modern memory models like FSRS.

Project HSK Nest
Status
Last Updated July 2026
Key Tags Next.js, TypeScript, Tailwind

Live Demo & Source

Experience the App

A fully functional instance is running at hsknest.com. You can try it instantly via Guest Mode or create an account to start studying.

HSK Nest Dashboard

The focus-ring dashboard tracking daily progress, due reviews, and an upcoming study forecast.

Overview

HSK Nest helps you remember vocabulary the efficient way. Using modern memory-model scheduling (FSRS), the app brings each word back right before you’d forget it. It ships pre-loaded with the full New HSK 3.0 (2021) vocabulary—levels 1 through 9—along with 3,000 real example sentences featuring pinyin, translation, and on-device pronunciation.

Key Features

  • Multi-language by design: The underlying data model treats every word as a combination of term, translation, and optional phonetic and metadata. This means any language fits perfectly.
  • Tunable scheduling algorithms: Choose between FSRS (modern memory-model scheduler), SM-2 (adaptive SuperMemo 2), or Leitner (5 fixed boxes).
  • Gesture-first study mode: A distraction-free, dark focus mode with a swipeable card stack.
  • Multiple practice modes: Go beyond flashcards with meaning quizzes, reading quizzes (symbol → sound), matching rounds, and sentence practice to recognize vocabulary in context.
  • HSK-level onboarding: Select your target HSK level at signup and the matching deck is automatically enrolled.
  • Self-Hostable and Private: Run it on your own server, keep everything on-device (even text-to-speech), and never depend on a cloud service.
HSK Nest Study Deck

Gesture-first flashcard study mode with on-device pronunciation and integrated example sentences.

Technical Foundation

Built on a robust, modern stack designed for speed and simplicity in deployment:

  • Next.js (App Router) & TypeScript
  • Tailwind CSS with shadcn/ui primitives and framer-motion for fluid interactions.
  • Prisma ORM (SQLite for dev, Postgres for prod)
  • NextAuth.js for secure, credentials-based authentication.

Every list shows how many words are in your queue, their recall strength, and what’s due now. The system gracefully handles duplicate words across lists so your study schedule never doubles up.

← All Projects Markus Berlit