English · 简体中文

Studify 🎓 — AI Tutoring & Learning Platform

Studify is a modern, AI-powered platform that unifies Courses, Classrooms, and Community with built-in Payments, Push Notifications, and an advanced AI layer (semantic search, learning paths, video transcription + Q&A, recommendations). Built with Next.js 15, Supabase, Redis, Stripe, OneSignal, QStash, LangChain, and HuggingFace.

Student Lifestyle

Goal:

Many students struggle with academic challenges, including difficulty understanding course material, limited access to personalized help, and inefficient study habits. Traditional tutoring services are often expensive, inconvenient, or unavailable outside fixed hours, leaving students without adequate support when they need it most. Develop a tutoring platform that provides on-demand, affordable, and personalized academic assistance. The platform connects students with tutors, offers interactive learning tools, and provides curated study resources to improve academic outcomes and boost student confidence.

Features · Tech Stack · Screenshots · Modules · PWA & Capacitor

Features

• Courses
  • Modules and Lessons (video/live/document/quiz/assignment/whiteboard)
  • Enrollment, Progress Tracking, Notes, Quizzes (auto grading), Certificates
  • Payments with Stripe (Checkout + Webhooks), free/paid courses
• Classrooms
  • Class creation, member roles (owner/tutor/student), Join via code
  • Assignments + Submissions + Grading, Live sessions + Attendance
  • Whiteboard, Recordings, Learning Path integration
• Community
  • Groups (public/private), Posts, Comments (nested), Reactions
  • Achievements, Check-ins, Hashtags, Recommendations
• AI Layer
  • Dual-embedding search (E5-Small + BGE-M3), RAG Q&A, Learning Coach
  • Video transcription (Whisper) with resilient queue + warmup
  • Personalized learning paths with Mermaid visualization
• Engagement & Growth
  • OneSignal push notifications with DB persistence
  • Gamification (points, achievements, learning stats)
• i18n & UX
  • next-intl multi-language, glass-morphism UI, dark/light themes

Tech Stack

• Frontend: Next.js 15 (App Router), React 19, TypeScript, TailwindCSS, shadcn/ui, Framer Motion
• Backend: Supabase (Auth/DB/Storage), Upstash Redis, QStash
• AI: LangChain, OpenRouter, HuggingFace Whisper, Mermaid
• Payments: Stripe (Checkout, Connect, Webhooks)
• Notifications: OneSignal
• i18n: next-intl

Screenshots

🎥 Video Demo

https://www.youtube.com/watch?v=gbdpsgFKaUI

Modules Overview

• Course (app/api/course/, db tables: course*, course_lesson, course_enrollment, course_progress, course_quiz_*)
  • Students: enroll, learn, track progress, take quizzes, write notes, pay
  • Tutors: manage course content, modules/lessons, analytics
  • Admin: approve/reject, moderation, analytics
• Classroom (app/api/classroom/, tables: classroom*)
  • Owner/Tutor: create class, manage members, assignments, live sessions
  • Students: join via code, submit assignments, attend live sessions
• Community (app/api/community/, tables: community_*)
  • Groups (public/private), posts, comments, reactions
  • Achievements, check-ins, hashtags, quizzes

Role-based access is enforced in APIs via utils/auth/server-guard.ts using authorize('student'|'tutor'|'admin') and DB-level ownership checks.

Highlights

• AI-first learning experience
  • Time-aware video Q&A with ±30s window, terminology cards, and jump-to-timestamp
  • Components and APIs: app/api/video/qa, components/ai-coach/*, components/ai/*, documentation/ai/AI_LEARNING_COACH_SYSTEM.md
• Rich video interactions
  • Endpoints: app/api/video/views, app/api/video/likes, app/api/video/danmaku, app/api/video/comments, app/api/video/comments/[id]/likes
  • Player: components/video/bilibili-video-player.tsx
• Dual-embedding semantic search
  • E5-Small (384d) + BGE-M3 (1024d) hybrid search, background queue, admin controls
  • Key files: db/migrations/20250905_embedding_system.sql, lib/langChain/embedding.ts, lib/langChain/vectorstore.ts, app/api/embeddings/*
• Background processing with resilience
  • QStash-powered pipelines with warmup and retry for Whisper transcription
  • Transcription step: app/api/video-processing/steps/transcribe/route.ts
• Payments and real payouts
  • Stripe Checkout + Connect with 90/10 split, scheduled earnings release
  • Endpoints: app/api/course/webhook, app/api/tutor/stripe-connect, app/api/tutor/earnings, app/api/tutor/earnings/release
• Notifications and growth
  • OneSignal push + DB persistence, unread counts, user preferences
  • Worker: public/OneSignalSDKWorker.js, hooks/components under components/ and hooks/
• Internationalization and UX
  • next-intl multi-language, glass-morphism design, dark/light themes, accessibility
• Admin and RBAC
  • Admin dashboards and APIs: app/api/admin/*, components/admin/*, hooks/admin/*

Performance and Security

• Security: RLS policies on all sensitive tables, API-level authorize(...), ownership checks, audit logs
• Reliability: React Query caching, Redis caching, warmup + retries, rate limits (e.g., danmaku), DB indexes and triggers
• Scalability: Background processors for embeddings/transcriptions, paginated APIs, hybrid vector search

PWA & Capacitor

• PWA layer (Serwist + Manifest)

  • Service Worker: defined in app/sw.ts, bundled via @serwist/next as public/sw.js (see next.config.ts). It is disabled in development and enabled only in production.
  • Manifest: public/manifest.json provides PWA metadata and icons. Ensure icons match your branding and sizes.
  • OneSignal worker: public/OneSignalSDKWorker.js is intentionally excluded from Serwist (see exclude in next.config.ts) and served with proper headers.
  • Verify: build and run locally, then check Chrome DevTools → Application → Service Workers.

    npm run build && npm start

• Capacitor mobile shell

  • Config: capacitor.config.ts

    const config: CapacitorConfig = {
      appId: "com.studify.platform.vercel.app",
      appName: "Studify",
      webDir: "empty", // using remote server
      server: { url: "https://studify-platform.vercel.app", cleartext: true },
    };

  • Development (Android/iOS): point server.url to your dev host
    • Android emulator: http://10.0.2.2:3000
    • Real device (same LAN): http://<LAN-IP>:3000
    • Steps:

      # 1) Start Next.js
      npm run dev
      
      # 2) Sync native project
      npx cap sync android   # or: npx cap sync ios
      
      # 3) Run the app
      npx cap run android    # or open with Android Studio: npx cap open android

    • Note: cleartext: true allows http during local testing; use https in production.
    • Deploy the web (e.g., to Vercel) and set server.url to the https domain (already set in this repo).
    • Then package the app:

      npx cap sync android && npx cap open android   # build/sign in Android Studio
      npx cap sync ios && npx cap open ios           # archive/sign in Xcode

    • Alternative (advanced): set webDir to a static export (e.g., out) if your app is purely static. SSR/App Router pages generally should keep server.url instead.
  • Push notifications on mobile
    • Web push continues to use public/OneSignalSDKWorker.js.
    • If moving to native push, integrate the respective Capacitor plugin and platform setup; this repo currently focuses on web push (OneSignal) with correct SW headers in next.config.ts.

Feedback

• User Feedback Form: https://forms.gle/cZSzsLVbRqysDcER9
• Issue Reporting Form: https://forms.gle/8ZYapQ9e9MFZ7EC97