Wachsen

Wachsen: Plan schedules, execute exams, scan physical questions, generate AI flashcards, & track revisions in one app. Powered By ChatGPT

Comment

Inspiration :

Standardized test preparation is often a lonely, static, and stressful journey. Students struggle with two major challenges: lack of clear, structured guidance (knowing what to study day-by-day) and isolated learning (missing the competitive and collaborative nature of studying with friends). And sometimes students want to give test one some specific topic for that there is tool to give that .

We built Wachsen (German for "to grow") to transform exam preparation from a passive chore into an active, personalized, and socially engaging roadmap to academic success.

What it does Wachsen AI-driven adaptive learning platform that generates complete study roadmaps, hosts real-time friends challenges, and reviews exam performance with an integrated AI Tutor.

Custom Mock Exams: Supports diverse formats including Multiple Choice, True/False, Integer-type, and Long Answer Questions (LAQ), with customizable timers and negative marking rules. Structured Study Roadmaps: Instantly generates Month-by-Month and Day-by-Day roadmaps, complete with a detailed Syllabus Checklist showing exact progress percentages. Social Challenges: Allows users to add friends and send "Exam Challenges." Friends can sit the same exam under identical conditions, comparing their scoring and answer keys side-by-side. Adaptive Revision Logs: Tracks every incorrect or skipped question, letting users retry their mistakes and automatically generates conceptual theory flashcards to master weak areas. Print Out : Students can print out the question paper from right there . Import Exams : Import freind's Public Exams and take test on that . AI Tutor & Mentor: Integrated directly into performance reviews and daily schedules, offering explanation of concepts, personalized grading on long-answer responses, and customized mock test guidance.

How we built it :

Frontend: React (TypeScript), styled with a highly customized Vanilla CSS and Tailwind CSS design system featuring fluid responsive typography, dynamic transitions, and a premium light/dark theme engine. State & Caching: TanStack Query (React Query) configured with smart caching policies, integrated with a local IndexedDB engine to store heavy JSON files (like revision data, exam logs) locally.

==> The App is that much optimized that you will feel it so smooth after you open it once .

This guarantees instant, zero-latency transitions and cold-start support.

Database & Auth:

Supabase handles real-time authentication, user profiles, and transactional data (exams, results, challenges, friendships).

Math Rendering:

Custom LaTeX parsing components (MathText) to correctly render inline and block formulas using mathematical notations. It can render most complex mathematical expressions very easily.

Challenges we ran into

  1. Synchronous State Transitions: Users experienced brief loading flashes when transitioning back and forth between lists and detailed reports. Even though data was cached inside IndexedDB, accessing it asynchronously caused a micro-delay. We resolved this by nesting TanStack Query on top of IndexedDB serving cached data synchronously in memory on session visits.

  2. Real-time Friendship Syncing: Cache invalidation issues meant that when one user accepted a friend request, the other user's list wouldn't update instantly. We configured real-time fetching properties on the Friends/Challenges queries to query Supabase dynamically on tab toggle.

3.Fully Formatted LaTeX Integration: ( THE TOUGHEST ONE ) As we are using the LaTex and in json , we faced a major issue , Json take backslash as a special escape character . so we have to tell the AI to generate double backslashes instead of one but AI also sometimes put one , as the AI is mostly trianed with one backslashes . so we added multiple layers , so currently Near 100% percent of the exams the AI generate is perfect.

Accomplishments that we're proud of :

  1. Not "Vibe Coded" Slop:
    We resisted the urge to just let LLMs write random code and hope it compiles. Instead, we wrote clean, structured React code . Yes we Used AI , but not blindly.

  2. Zero-Flash Cached Navigation: Reaching a point where heavy analytics charts, results, and revision logs load instantly with absolutely zero loading spinners on repeat visits.

  3. Fully Formatted LaTeX Integration: Making sure mathematical questions, subjects, and study plan steps render mathematical equations perfectly across all checklists and question displays.

  4. Seamless Mobile Styling: Designing complex elements like double-nested grids, large analytics charts, and tabular friend panels to look premium and scale naturally across viewports.

  5. All in one app.

What we learned :

  1. State Hydration Balance: We learned the power of bridging in memory state management (React Query) with persistent client side storage (IndexedDB) to design offline-capable, responsive apps.

2.Tailwind Constraints: Discovered that relying on non-standard intermediate styling units can break production builds, leading us to build clean extension scales inside tailwind.config.js.

What's next for Wachsen :

  1. Collaborative Live Study Rooms: Real time group mock tests where friends can take a test together synchronously with active leaderboards. with camera on .

  2. Offline-First Mock Sessions: Complete local caching of questions and assets so users can sit entire mock tests offline, with automatic syncing once internet connection is restored.

  3. Deep AI Weakness Mapping: An automated visual heat map of subjects showing which micro-concepts a student consistently gets wrong, along with automated daily booster tasks for those exact sub-topics.

  4. Make a webview app and launch on play / app stores .

Built With

  • avatar
  • express.js
  • indexed
  • lucide-icons
  • mathjax
  • react
  • recharts
  • schema-validation
  • storage
  • supabase-(postgres-tables
  • tailwind-css
  • tanstack-query
  • typescript
  • vite
Share this project:

Updates