Bookly AI

Bookly AI is an AI-powered study assistant that lets you upload PDFs, chat with retrieval-augmented answers, generate quizzes, review progress, and explore related YouTube content. It uses Supabase for authentication and persistence, Google Gemini for generation, and a lightweight RAG pipeline for grounded responses.

Highlights

• Chat with RAG: Ask questions; answers cite relevant PDF pages when available.
• PDF ingestion: Upload, parse, embed, and search your documents.
• Quizzes and progress: Generate quizzes, track attempts, view weaknesses and dashboard.
• YouTube suggestions: Surface related topics and recommended videos.
• Auth & sessions: Supabase auth with middleware-protected routes.

User Flow

graph TD
    Start([Student Opens App]) --> Auth{Authenticated?}
    Auth -->|No| Login[Login/Signup Page]
    Login --> Dashboard
    Auth -->|Yes| Dashboard[Main Dashboard]
    
    Dashboard --> LeftPanel[Left Sidebar]
    Dashboard --> CenterPanel[Center Chat Area]
    Dashboard --> RightPanel[Right PDF Preview]
    
    %% Left Sidebar Flow
    LeftPanel --> NewChat[New Chat Button]
    LeftPanel --> UploadedFiles[Uploaded Files Section]
    LeftPanel --> ProgressDash[Progress Dashboard]
    LeftPanel --> ChatHistory[Chat History List]
    
    NewChat --> CreateNewChat[Create New Chat]
    CreateNewChat --> CenterPanel
    
    UploadedFiles --> ViewFiles[View All PDFs]
    ViewFiles --> SelectPDF[Select PDF]
    SelectPDF --> LoadPDF[Load PDF in Right Panel]
    SelectPDF --> SetContext[Set as Chat Context]
    
    ProgressDash --> ViewStats[View Statistics]
    ViewStats --> QuizScores[Quiz Scores & History]
    ViewStats --> Strengths[Strength Areas]
    ViewStats --> Weaknesses[Weak Areas]
    ViewStats --> AIAnalysis[AI-Generated Analysis]
    ViewStats --> Activity[Activity Timeline]
    
    ChatHistory --> SwitchChat[Click Chat]
    SwitchChat --> LoadChat[Load Chat Context]
    LoadChat --> CenterPanel
    
    %% Center Chat Area Flow
    CenterPanel --> InputBox[Chat Input Box]
    InputBox --> TextInput{Input Type}
    TextInput -->|Text| TypeMessage[Type Question/Message]
    TextInput -->|Voice| VoiceInput[Voice Input]
    TextInput -->|PDF Upload| UploadPDF[Upload New PDF]
    
    TypeMessage --> SendMessage[Send Message]
    VoiceInput --> TranscribeVoice[Transcribe Audio]
    TranscribeVoice --> SendMessage
    
    UploadPDF --> ProcessPDF[Process & Store PDF]
    ProcessPDF --> ExtractText[Extract Text Content]
    ExtractText --> ChunkEmbed[Chunk & Embed Text]
    ChunkEmbed --> StoreVectors[Store in pgvector]
    StoreVectors --> ShowUpload[Show in Uploaded Files]
    ShowUpload --> LoadPDF
    
    SendMessage --> QuizOption{Quiz Generation?}
    QuizOption -->|Checkbox Selected| GenerateQuiz[Generate Quiz Immediately]
    QuizOption -->|Not Selected| NormalChat[Process as Chat Message]
    
    NormalChat --> RAGQuery[RAG Query Process]
    RAGQuery --> EmbedQuery[Embed User Query]
    EmbedQuery --> VectorSearch[Search Similar Chunks]
    VectorSearch --> RetrieveContext[Retrieve Relevant Context]
    RetrieveContext --> LLMResponse[Generate LLM Response]
    LLMResponse --> CitedAnswer[Answer with Citations]
    CitedAnswer --> DisplayAnswer[Display in Chat]
    DisplayAnswer --> ShowCitations[Show Page Numbers & Quotes]
    
    %% Quiz Generation Flow
    GenerateQuiz --> QuizType[Select Quiz Type]
    QuizType --> MCQ[MCQ Generator]
    QuizType --> SAQ[SAQ Generator]
    QuizType --> LAQ[LAQ Generator]
    QuizType --> Mixed[Mixed Quiz]
    
    MCQ --> GenMCQ[Generate MCQ Questions]
    SAQ --> GenSAQ[Generate SAQ Questions]
    LAQ --> GenLAQ[Generate LAQ Questions]
    Mixed --> GenMixed[Generate Mixed Questions]
    
    GenMCQ --> RenderQuiz[Render Quiz UI]
    GenSAQ --> RenderQuiz
    GenLAQ --> RenderQuiz
    GenMixed --> RenderQuiz
    
    RenderQuiz --> TakeQuiz[Student Takes Quiz]
    TakeQuiz --> AnswerQs[Answer Questions]
    AnswerQs --> SubmitQuiz[Submit Quiz]
    
    SubmitQuiz --> EvaluateQuiz[Evaluate Answers]
    EvaluateQuiz --> MCQScore[Auto-Score MCQs]
    EvaluateQuiz --> AIScore[AI Score SAQs/LAQs]
    
    MCQScore --> CalcScore[Calculate Total Score]
    AIScore --> CalcScore
    CalcScore --> ShowResults[Show Results]
    
    ShowResults --> DetailedFeedback[Detailed Feedback]
    DetailedFeedback --> CorrectAnswers[Show Correct Answers]
    DetailedFeedback --> Explanations[Provide Explanations]
    DetailedFeedback --> ConceptReview[Concept Review Links]
    
    ShowResults --> StoreAttempt[Store Quiz Attempt]
    StoreAttempt --> UpdateProgress[Update Progress Data]
    UpdateProgress --> AnalyzePerformance[Analyze Performance]
    AnalyzePerformance --> IdentifyWeakness[Identify Weak Topics]
    AnalyzePerformance --> IdentifyStrength[Identify Strong Topics]
    AnalyzePerformance --> UpdateDashboard[Update Dashboard Stats]
    
    ShowResults --> RegenerateOption{Regenerate Quiz?}
    RegenerateOption -->|Yes| GenerateQuiz
    RegenerateOption -->|No| ContinueChat[Continue Chat Session]
    
    %% Additional Features
    DisplayAnswer --> YouTubeRec[YouTube Recommendations]
    YouTubeRec --> FetchVideos[Fetch Related Videos]
    FetchVideos --> DisplayVideos[Display Video Cards]
    DisplayVideos --> WatchVideo[Student Watches Video]
    
    %% Right Panel Flow
    RightPanel --> PDFDisplay[PDF Display Component]
    PDFDisplay --> TogglePDF{Toggle Visibility}
    TogglePDF -->|Hide| CollapsePDF[Collapse Panel]
    TogglePDF -->|Show| ExpandPDF[Expand Panel]
    
    PDFDisplay --> ResizePanel[Resize Panel Width]
    ResizePanel --> DragResize[Drag to Resize]
    
    PDFDisplay --> NavigatePDF[Navigate PDF]
    NavigatePDF --> PageControls[Page Navigation]
    NavigatePDF --> SearchPDF[Search in PDF]
    NavigatePDF --> JumpToPage[Jump to Cited Page]
    
    ShowCitations -.->|Click Citation| JumpToPage
    
    ContinueChat --> InputBox
    WatchVideo --> ContinueChat
    
    UpdateDashboard --> ProgressDash
    
    style Start fill:#e1f5e1
    style Dashboard fill:#e3f2fd
    style GenerateQuiz fill:#fff3e0
    style ShowResults fill:#f3e5f5
    style StoreAttempt fill:#fce4ec

Loading

Tech Stack

• Framework: Next.js 15 (App Router), React 19
• Styling/UI: Tailwind CSS v4, shadcn/ui components, Radix UI primitives, lucide-react
• State: Zustand
• Auth/DB: Supabase (@supabase/ssr, @supabase/supabase-js)
• LLM: Google Gemini (@google/generative-ai)
• PDF: pdf-parse, @react-pdf-viewer/*

Monorepo structure (key paths)

• app/ — Next.js routes (App Router) and API routes under app/api/*
• components/ — UI components (layout, chat, files, quiz, progress)
• lib/ — Server/client utilities: auth, env, RAG, quiz, Supabase clients
• store/ — Zustand stores (auth, chat, pdf, quiz, ui)
• db/migrations/ — SQL migrations for Supabase schema and features

---

Local Development Setup

0) Prerequisites

• Node.js 18+ (LTS recommended)
• npm (comes with Node) or yarn/pnpm
• A Supabase project (free tier is fine)
• API keys for Google Gemini and optionally YouTube Data API v3

1) Get the code

# Clone the repository and enter the project
git clone https://github.com/<your-org>/<your-repo>.git
cd <your-repo>/bookly

2) Create and configure Supabase

1. Create a new project at https://supabase.com.
2. In your project settings, locate:
   • Project URL (Supabase URL)
   • anon public key (Client)
   • service_role key (Server; keep private, server-only)
3. Apply the SQL migrations from /testDB/dbSchema.sql using the Supabase SQL Editor.

Note: Some features (RAG, quizzes, recommendations) require the later migrations.

3) Configure environment variables

Create a .env.local file in bookly/ with:

# Google Gemini (required for chat and quizzes)
GEMINI_API=your_gemini_api_key

# YouTube Data API (optional; enables recommendations)
YT_API_KEY=your_youtube_api_key

# Server-side Supabase (service role used by API routes)
SUPABASE_URL=https://your-project-id.supabase.co
SUPABASE_ANON_API_KEY=your_anon_public_key
SUPABASE_SERVICE_KEY=your_service_role_key

# Public (browser) Supabase
NEXT_PUBLIC_SUPABASE_URL=https://your-project-id.supabase.co
NEXT_PUBLIC_SUPABASE_ANON_KEY=your_anon_public_key

Environment resolution:

• Server utilities use SUPABASE_URL, SUPABASE_ANON_API_KEY, SUPABASE_SERVICE_KEY.
• Browser client prefers NEXT_PUBLIC_* and falls back to server vars in dev.

4) Install dependencies

npm install

5) Run the dev server

npm run dev

Open http://localhost:3000.

---

How Authentication Works

• Sign in/up pages: /signin, /signup.
• Middleware (middleware.ts) allows public paths and protects the rest by checking Supabase session cookies.
• Server APIs use @supabase/ssr to read/write session cookies and get the user.

If auth is not configured, protected routes will redirect to /signin.

---

Core Features & Flows

Feature Checklist (from ProjectPlan)

• Source selector: It has been implemented to allow choosing across all uploaded PDFs or a specific PDF, with file uploads supported. The upload flow and file listing are provided via app/upload/page.tsx, app/files/page.tsx, and the components/files/* and components/shell/* panels.

• PDF viewer: It has been implemented to display the selected PDF alongside the chat. The right panel provides page navigation, search, and resizing controls (see components/shell/pdf/* and the resizable layout in components/ui/resizable.tsx).

• Quiz Generator Engine (MCQ, SAQ, LAQ): It has been implemented to generate and render quizzes, capture answers, score submissions, store attempts, and provide explanations. Generation and evaluation routes exist under app/api/quiz/*, with UI in components/quiz/* and progress storage in quiz_attempts/answers tables.

• Progress tracking: It has been implemented to persist strengths/weaknesses and surface a dashboard. APIs live under app/api/progress/*, UI under components/progress/*, and storage in user_progress and user_weaknesses tables.

• Chat UI (ChatGPT-inspired): It has been implemented with a left drawer for chats, a main chat area, and an input box, supporting new chat creation and switching. The layout is responsive across devices (see components/shell/* and components/layout/MainLayout.tsx).

• RAG answers with citations: It has been implemented with PDF ingestion (chunk + embed), similarity search, and citation surfaces in responses. The pipeline lives in lib/rag.ts and related API routes (app/api/pdf/*, app/api/chat), returning cited page numbers and quotes when available.

• YouTube Videos Recommender: It has been implemented to recommend educational videos relevant to context and detected weaknesses. Endpoints are under app/api/youtube/*, with UI in components/youtube/* and persistence in youtube_recommendations.

Operational Endpoints and Flows

• Chat with Retrieval: POST /api/chat — Session is validated, user provisioned, optional rate limiting applied. When a pdfId is provided and ready, the query is embedded, top chunks are retrieved, and a response is streamed from Gemini with citations; messages are persisted in messages.
• Messages History: GET /api/messages?chatId=... — Chat ownership is enforced and ordered messages are returned.
• PDFs: Uploading, parsing, embedding, and search are performed via app/api/pdf/*, with status tracked in pdfs and chunk vectors in chunks.
• Quizzes & Progress: Quiz generation/evaluation under app/api/quiz/*; dashboard and history powered by app/api/progress/* and visualized in components/progress/*.
• YouTube Recommendations: Recommendation APIs live under app/api/youtube/* and are driven by content/topics and weaknesses, gated by YT_API_KEY.

---

Scripts

• npm run dev — Start Next.js dev server
• npm run build — Build for production
• npm run start — Start production server
• npm run lint — Run ESLint

---

Troubleshooting

• "Authentication required" on protected pages:

  • Confirm .env.local has valid NEXT_PUBLIC_SUPABASE_URL and NEXT_PUBLIC_SUPABASE_ANON_KEY.
  • Ensure your Supabase Auth providers are enabled and redirect URLs include http://localhost:3000.

• Chat returns 500 "GEMINI_API missing":

  • Set GEMINI_API in .env.local.

• Missing data or SQL errors:

  • Verify all migrations in db/migrations/ were applied in order.

• Windows or local TLS/cookie oddities:

  • The middleware and server helpers will try to decode the Supabase access token from cookies as a fallback.
  • Make sure third-party cookies are not blocked in the browser during local testing.

• PDF viewer issues (canvas/pdfjs):

  • The app stubs canvas in next.config.ts; ensure you’re not bundling canvas on the client.

---

Security Notes

• Never expose SUPABASE_SERVICE_KEY to the browser or commit it to VCS.
• Use separate environment files or secrets for production vs. development.

---

License

Proprietary. All rights reserved.