StudyFlow — NotebookLM × Notion Learning System

An intelligent study system that bridges NotebookLM (knowledge synthesis) and Notion (structured, persistent notes) — orchestrated by Claude.

StudyFlow transforms the way CSE students learn. Upload any resource to NotebookLM — PDFs, YouTube lectures, web articles, research papers — and StudyFlow automatically structures the content into Cornell-style notes, spaced-repetition flashcards, Feynman explanations, and active recall quizzes inside Notion.

Why This Pairing Is Powerful

Tool	Role
NotebookLM	Synthesizes uploaded material into insights, Q&A, mind maps, quizzes, podcasts
Notion	Permanent, structured, beautiful knowledge base with spaced repetition
Claude	Intelligent orchestrator bridging the two — via a /studyflow skill and 9 Claude Desktop tools

The result is a system better than any single tool: NotebookLM is excellent at synthesizing uploaded material; Notion is excellent at structured, persistent, linked knowledge bases; Claude is the intelligent glue.

Features

• 📝 Structured Notes — Cornell format with learning objectives, key definitions, code examples, Feynman explanations
• 🧠 SM-2 Spaced Repetition — Flashcards with Ease Factor × Interval scheduling (New → Learning → Review → Mastered)
• 🔍 Active Recall — 5–8 auto-generated questions on every note page
• 🧪 Feynman Technique — ELI12 explanations that surface knowledge gaps
• ❓ Quiz Bank — Import NotebookLM quizzes with accuracy tracking
• 🎧 Podcast Notes — Extract key insights from NotebookLM Deep Dive audio
• 🗺️ Mind Map Import — Convert mind maps to Notion page hierarchies
• 📅 Exam Prep Mode — Calculates daily review targets, crunch schedules
• 🔍 Multi-Source Finder — Search YouTube + DuckDuckGo + arXiv in parallel, no API keys
• 🤖 9 Claude Desktop MCP Tools — Source discovery and notebook management

Architecture

/studyflow <command> <notebook> [options]
      │
      ├── notebooklm CLI ──── query, generate artifacts, extract quizzes/mind maps/podcasts
      │
      ├── Notion MCP ───────── create pages, databases, update SM-2 flashcard intervals
      │
      └── studyflow_mcp_server.py ── 9 Claude Desktop tools for source ingestion
                                     (YouTube transcripts, web scraping, arXiv, DuckDuckGo)

Quick Start

# 1. Install dependencies
pip install -r requirements.txt
pip install yt-dlp

# 2. Register MCP server with Claude Desktop
# See docs/claude-desktop-setup.md

# 3. Copy SKILL.md to Claude Code skills directory
cp SKILL.md ~/.claude/skills/studyflow/SKILL.md

# 4. Create Notion workspace
# In Claude Code:
/studyflow setup

# 5. Start studying
/studyflow notes "Your Notebook Name" --topic "Dynamic Programming"

See docs/SETUP.md for complete installation instructions.

Claude Code Commands

Command	Description
/studyflow setup	Create Notion workspace (one-time)
/studyflow notes <nb> [--topic T] [--type concept|algo|tech]	Generate structured notes
/studyflow flashcards <nb> [--count N] [--cloze]	Generate SM-2 flashcards
/studyflow review	Process today's due cards (SM-2)
/studyflow quiz <nb>	Import NotebookLM quiz into Quiz Bank
/studyflow feynman <nb> --topic T	Generate Feynman explanation
/studyflow study <nb>	Full study session (notes + cards + quiz)
/studyflow exam <course> --days N	Exam prep mode
/studyflow brief	Daily study brief
/studyflow search-youtube <query> [--filter lecture|tutorial|short]	Find YouTube videos
/studyflow add <nb> <url>	Add source to notebook
/studyflow podcast <nb>	Extract podcast key insights
/studyflow mindmap <nb>	Import mind map as page hierarchy
/studyflow slides <nb>	Import slide deck summaries
/studyflow project <name>	Create project + curriculum links
/studyflow sync <nb>	Re-sync updated notebook

Claude Desktop MCP Tools

Tool	Description
studyflow_list_notebooks	List all NotebookLM notebooks
studyflow_find_sources	Search YouTube + Web + arXiv in parallel
studyflow_search_youtube	YouTube search with lecture/tutorial/short filter modes
studyflow_add_youtube	Extract VTT transcript → upload to NotebookLM
studyflow_add_web	Site-specific scraping → upload to NotebookLM
studyflow_add_url	Add URL directly to NotebookLM
studyflow_add_research	NotebookLM web research (fast/deep mode)
studyflow_ask	Query a notebook with a question
studyflow_source_list	List sources in a notebook

Notion Workspace Structure

📚 StudyFlow
├── 📈 Dashboard           (command reference, daily stats)
├── 🗂️ Courses             (one sub-page per notebook/course)
├── 🔧 Projects & Freelance
├── 📚 Knowledge Base      (all notes database)
├── 🃏 Flashcard Vault     (SM-2 flashcards database)
├── ❓ Quiz Bank           (imported quizzes database)
└── 📅 Study Sessions      (session log database)

SM-2 Algorithm

StudyFlow implements the classic SM-2 spaced repetition algorithm:

Again: interval=1, repetitions=0
Hard:  interval=max(1, round(interval × 1.2)), ease_factor=max(1.3, ef - 0.15)
Easy:  interval = 1 | 6 | round(interval × ease_factor)  [based on repetition count]
       ease_factor = min(4.0, ease_factor + 0.1)

Status: reps=0 → 🆕 New | 1-2 → 📖 Learning | 3-5 → 🔄 Review | 6+ & interval>21 → ✅ Mastered

Target User

CSE undergrad studying AI, Compilers, Theory of Computing, IoT, DSA, OS, Computer Architecture — also freelancing and self-learning new tech stacks on the job.

Requirements

• Python 3.11+
• notebooklm CLI — authenticated (notebooklm login)
• Claude Code with Notion MCP configured
• yt-dlp (optional — for YouTube transcript extraction)

File Structure

studyflow/
├── studyflow_mcp_server.py      # Claude Desktop MCP server (9 tools)
├── SKILL.md                     # Claude Code skill definition (16 commands)
├── requirements.txt             # Python dependencies
├── docs/
│   ├── SETUP.md                 # Full installation guide
│   ├── claude-desktop-setup.md  # MCP server registration
│   └── notion-schema.md         # Database property schemas
└── examples/
    └── claude_desktop_config.example.json

License

MIT