The extension helps students prepare for exams by processing lecture PDF slides in a two-stage pipeline:

1. Preprocessing: Clean and extract content from PDFs
2. Analysis: Generate structured study notes with key topics and explanations

sudo apt-get update
sudo apt-get install -y npm nodejs uv git

npm install -g @qwen-code/qwen-code@latest
qwen --version

Optional: Use OpenRouter

Edit ~/.qwen/.env:

OPENAI_API_KEY=your_openrouter_api_key
OPENAI_BASE_URL=https://openrouter.ai/api/v1
OPENAI_MODEL=x-ai/grok-4-fast

Edit ~/.qwen/settings.json:

{
  "selectedAuthType": "openai"
}

Get your API key: https://openrouter.ai/keys Browse models: https://openrouter.ai/models

npm install -g @google/gemini-cli@latest
gemini --version

For Qwen:

qwen extensions install https://github.com/planetis-m/study-ai.git

For Gemini:

gemini extensions install https://github.com/planetis-m/study-ai.git

qwen

gemini

Then run /help to confirm that the study-ai commands are available.

OCR support relies on Tesseract. Install it with:

sudo apt install -y tesseract-ocr tesseract-ocr-eng

Then set the environment variable so the extension can find the language data:

export TESSDATA_PREFIX=/usr/share/tesseract/tessdata

After that, OCR will work automatically when you use commands like:

/study-ai:study-notes scanned-lecture.pdf --use-ocr yes

Your <workspace> is the folder where you store your lecture slides (for example, ~/Documents/COMP417).

cd <workspace>
# Clean a single PDF
/study-ai:clean lecture1.pdf

# Clean with OCR reprocessing for incomplete pages
/study-ai:clean-ocr lecture1.pdf

# Clean specific pages
/study-ai:clean lecture1.pdf --start-page 5 --end-page 15

# Analyze already-cleaned text content
/study-ai:analyze "Paste your cleaned slide content here..."

# Complete pipeline (extract, clean, analyze)
/study-ai:study-notes lecture1.pdf

# Redirect output to a file
/study-ai:study-notes lecture1.pdf > notes-lecture1.md

# Use OCR for scanned PDFs
/study-ai:study-notes scanned-lecture.pdf --use-ocr yes

# Generate practice quizzes from content or notes
/study-ai:quiz "Paste your slide content here..."

study-ai/
├── AGENTS.md            # Context file with assistant behavior
├── qwen-extension.json  # Extension config
└── commands/
    └── study-ai/
        ├── clean.toml           # Basic preprocessing command
        ├── clean-ocr.toml       # Deep cleaning with OCR reprocessing
        ├── analyze.toml         # Analysis command
        ├── study-notes.toml     # Complete pipeline command
        └── quiz.toml            # Quiz command

1. Separation of Concerns: Commands are separated into discrete stages (clean, analyze) while also providing a convenient combined command (study-notes)
2. Flexible Input: Users can either pipe PDF files directly or paste pre-extracted text for analysis
3. Context-Driven Behavior: The AGENTS.md file establishes the assistant's role and behavior patterns, which commands reinforce through their specific prompts
4. Integration with pdf-reader: Commands leverage the existing PDF tools (read_pdf_text, read_by_ocr) without reimplementing PDF parsing
5. Exam-Focused: All prompts emphasize exam preparation, key topics, and deep understanding rather than general summarization

This project is licensed under Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).