🚀 TRACKER: Master Development Roadmap

## 🟢 Phase 1: MVP (Minimum Viable Product)

### Core Features
- [ ] **Conversational Syllabus Generator**: Chat-based interface to define learning goals, target audience, and difficulty
- [ ] **Structured Course Planning**: Automated creation of Topics and Subtopics with strict learning outcomes
- [ ] **Smart Content Curation**: Search and filter YouTube videos based on metadata (views, likes, duration) vs. relevance
- [ ] **Course Sequencing**: AI logic to re-order videos linearly (Beginner → Advanced) rather than random lists
- [ ] **Playlist Automation**: One-click generation of actual YouTube Playlists on the user's account
- [ ] **Simple Video Player UI**: A clean interface to watch the curated course without YouTube distractions

---

## 🛠️ Infrastructure & Setup

- [ ] Initialize Git repository & .gitignore (Python/Node)
- [ ] Set up FastAPI backend structure (Folder layout)
- [ ] Set up React/Vite frontend structure
- [ ] Configure docker-compose.yml for Redis and PostgreSQL
- [ ] Create .env.example template for API keys
- [ ] Set up GitHub Actions for basic linting (Ruff/Black)

---

## 🧠 Backend: Core Logic (Python)

- [ ] Feature: Implement Pydantic Models for Course, Topic, Video
- [ ] Feature: Build GeminiService for Syllabus Generation (Prompt Engineering)
- [ ] Feature: Create API Endpoint: POST /generate-syllabus
- [ ] Feature: Implement YouTubeService using yt-dlp to fetch video metadata
- [ ] Feature: Build filtering algorithm (remove shorts, low-rated videos)
- [ ] Feature: Build PlaylistService to interact with official YouTube Data API

---

## 🤖 AI & RAG Layer (Phase 2)

- [ ] Feature: Implement Transcript Downloader service
- [ ] Feature: Set up Vector DB connection (Redis/Chroma)
- [ ] Feature: Create Embedding pipeline (Text → Vector)
- [ ] Feature: Create RAG Query Endpoint: POST /ask-tutor

---

## 🎨 Frontend: User Interface (React/Bun)

- [ ] UI: Build "Chat Interface" for the initial course negotiation
- [ ] UI: Build "Course Preview" card (Syllabus tree view)
- [ ] UI: Build "Video Player" page with side-by-side Chat Tutor
- [ ] Integration: Connect Frontend to Python Backend APIs

---

## 🟡 Phase 2: The AI Tutor (RAG)

- [ ] **Transcript Extraction**: Automated fetching of captions/transcripts via yt-dlp
- [ ] **Vector Database Integration**: Storing transcript chunks in Redis/ChromaDB
- [ ] **Contextual Q&A**: "Ask the video" feature where users can ask questions about the specific content they are watching

---

## 🔴 Phase 3: Advanced Features (Future)

- [ ] **Audio-to-Text Fallback**: Using Whisper for videos that don't have captions
- [ ] **Progress Tracking**: User accounts to track watched videos and course completion %
- [ ] **PDF/Quiz Generation**: AI generates a summary PDF or a quiz after every module

---

## 📋 Labels to Use

- `backend` - Python, FastAPI, Database logic
- `frontend` - React, Tailwind, UI/UX
- `ai-prompting` - Gemini prompts, System instructions
- `rag-pipeline` - Vector DB, Embeddings, Transcripts
- `youtube-integration` - yt-dlp, YouTube API quirks
- `priority: high` - MVP blockers