Read smarter. Think deeper. Learn measurably.
DeepReview is an intelligent web platform that helps students and instructors get more out of academic research papers. Instead of passively reading, students upload a paper and the platform extracts its structure, opens an AI chat grounded only in the article's content, and challenges them with a Socratic learning bot that adapts difficulty to their answers.
Behind the scenes, every interaction feeds a measurable picture of comprehension and critical thinking — which instructors can monitor through a dedicated analytics dashboard.
The goal isn't to summarize papers for students — it's to make them better, faster, more critical readers.
- ⏱️ Reduce reading time — surface a paper's key structure and ideas in seconds.
- 🧠 Strengthen critical thinking — emphasize reasoning over rote memorization.
- 📊 Make learning measurable — turn each session into trackable progress data.
- 👩🏫 Empower instructors — provide analytics, reports, and oversight tools.
| Role | What they do |
|---|---|
| 🎓 Students | Upload, analyze, discuss, and compare academic articles; learn through guided Socratic dialogue. |
| 👩🏫 Instructors | Track student progress, review comprehension & critical-thinking reports, and manage content. |
- 📄 Upload papers in PDF — text and page count are extracted automatically.
- 🔍 Automatic analysis with Google Gemini, extracting:
- Title · Authors · Abstract · Keywords · Publication year · Main topics
- 💬 Grounded academic chat — answers are based only on the uploaded article's content.
- 🤔 Socratic Learning Bot:
- Progressive, multi-level questions (difficulty levels 1–6)
- Understanding-oriented feedback on each answer
- Difficulty that adapts in real time to performance
- ⚖️ Article comparison — compare two papers by topic, methodology, and results.
- 📈 Personal progress tracking — scores, strengths, weaknesses, and recommendations.
- 📥 Export insights and sessions to PDF or Word.
- 🗂️ Administrative dashboard with student and content management.
- 👀 Progress monitoring across all students.
- 📊 Analytics & visualizations of comprehension and critical thinking.
- 📝 Per-student reports drilling into individual learning paths.
| Layer | Technologies |
|---|---|
| Frontend | Next.js 16 (App Router), React 19, TypeScript, Tailwind CSS |
| Backend | Next.js Server Actions & Route Handlers (Node runtime) |
| AI | Google Gemini (gemini-2.5-flash) — analysis, grounded chat, Socratic bot |
| Database & Auth | Supabase (PostgreSQL, Auth, Storage) |
| PDF & Export | pdf-parse (extraction) · jspdf, docx, html2canvas (export) |
| Data Viz | Recharts |
| Validation & Security | Zod schemas · custom HMAC-signed CAPTCHA |
| Deployment | Vercel |
DeepReview/
├── DeepReviewApp/ # Next.js application
│ ├── src/
│ │ ├── app/
│ │ │ ├── api/ # Route handlers: analyze · chat · socraticbot · upload
│ │ │ ├── auth/ # Login & register
│ │ │ └── dashboard/
│ │ │ ├── student/ # upload · chat · socraticbot · compare · library · profile
│ │ │ └── instructor/ # students · articles · analytics
│ │ ├── actions/ # Server actions (auth, articles, chat, instructor, profile…)
│ │ ├── components/ # UI (student / instructor / auth / shared / ui)
│ │ ├── lib/ # Supabase clients, Gemini, export, captcha
│ │ ├── types/ # Article, Message, SocraticMessage, StudentProgress
│ │ └── utils/ # Validation
│ └── public/
└── ARCH.jpeg # Architecture diagram
📐 See ARCH.jpeg for the full architecture diagram.
- Node.js 18+ and npm
- A Supabase project (with a Storage bucket named
articles) - A Google Gemini API key
| Command | Description |
|---|---|
npm run dev |
Start the development server |
npm run build |
Build for production |
npm run start |
Run the production build |
npm run lint |
Run ESLint |
- Upload — A student uploads a PDF.
pdf-parseextracts the text and page count, the file is stored in Supabase Storage, and an article record is created. - Analyze — Gemini extracts structured metadata (title, authors, abstract, keywords, topics) and saves it back to the database.
- Engage — The student chats with an AI grounded in the article, or starts a Socratic session with adaptive, multi-level questioning.
- Measure — Each session produces comprehension, critical-thinking, and quality scores, plus strengths, weaknesses, and recommendations.
- Monitor — Instructors review aggregated analytics and per-student reports.
- 🔗 Live App — https://deep-review-app.vercel.app
- 🧑💻 GitHub — https://github.com/Mohamedkhattab02/DeepReview.git
- 🎥 Demo Video — https://drive.google.com/file/d/1Gjpq9Lo93H9_spa3F-LLmNNhdkrh2mqx/view
| Name |
|---|
| Mohamed Khattab |
| Khaled Habashi |
| Samuel Zreik |