Canvas MCP Server

Overview

The Canvas MCP Server helps AI assistants like Claude access Canvas LMS content (assignments, syllabi, deadlines), allowing students to quickly ask questions like:

• "What's due this week?"
• "Send me tomorrow's notes."

This simplifies finding course info by automatically syncing Canvas data to your AI assistant.

Features

• Canvas Integration: Automatically sync courses, assignments, modules, announcements.
• Flexible Parsing: Extract syllabus details from PDFs, Word docs, HTML.
• Term Filtering: Only sync current-term courses.
• Opt-Out: Easily exclude courses you don't want indexed.

Quick Tech Overview

• Language: Python 3.12+
• Database: SQLite (local storage, zero config)
• Canvas API: canvasapi
• Document Parsing: pdfplumber (PDF), python-docx (Word), BeautifulSoup4 (HTML)
• AI Integration: Model Context Protocol (MCP)
• Package Management: UV (faster pip alternative)

Setup Instructions

Prerequisites

• Python 3.12+
• UV Package Manager
• Canvas API Key & URL
• Claude Desktop with MCP support (download)

Mac Installation

Install UV if needed:

curl -LsSf https://astral.sh/uv/install.sh | sh

Set up Canvas MCP:

git clone https://github.com/AdityaPrakash-26/Canvas-MCP.git && cd Canvas-MCP
uv venv --seed
source .venv/bin/activate
uv sync
echo "CANVAS_API_KEY=your_canvas_api_key_here" > .env
echo "CANVAS_API_URL=https://canvas.yourschool.edu" >> .env
uv run init_db.py

Configure Claude Desktop (Mac)

Open claude_desktop_config.json (Claude > Settings > Advanced) and add:

"Canvas MCP": {
  "command": "/path/to/uv",
  "args": [
    "run",
    "--directory",
    "/absolute/path/to/Canvas-MCP",
    "canvas-mcp"
  ]
}

Replace /path/to/uv with your UV executable path (which uv) and /absolute/path/to/Canvas-MCP with your cloned repository location.

Windows Installation (WSL)

Please use WSL.

After installing WSL:

1. Clone the repository into your WSL home:

git clone https://github.com/AdityaPrakash-26/Canvas-MCP.git ~/Canvas-MCP
cd ~/Canvas-MCP

2. Install UV and dependencies:

curl -LsSf https://astral.sh/uv/install.sh | sh
uv venv --seed
source .venv/bin/activate
uv sync
echo "CANVAS_API_KEY=your_canvas_api_key_here" > .env
echo "CANVAS_API_URL=https://canvas.yourschool.edu" >> .env
uv run python init_db.py

Configure Claude Desktop (Windows/WSL)

Edit Claude Desktop's claude_desktop_config.json with:

"Canvas MCP": {
  "command": "wsl.exe",
  "args": [
    "-d",
    "Ubuntu",
    "--exec",
    "/home/<USER>/.local/bin/uv", # REPLACE, run `which uv` to find ur `uv` install directory
    "run",
    "--directory",
    "/home/<USER>/Canvas-MCP", # REPLACE
    "canvas-mcp"
  ]
}

Replace <USER> with your WSL username. Also, verify your uv absolute path before copy pasting.

Development Commands

• Run the server in inspector mode so you can look at it:

uv run mcp dev --with . src/canvas_mcp/__main__.py

• Run tests:

uv run pytest

• Format/lint code:

uv run ruff format .
uv run ruff check . --fix

• Static type checking:

uv run mypy src

Troubleshooting Quick Tips

• Canvas API issues? Verify .env keys/URLs.
• Database issues? Reset:

rm -f data/canvas_mcp.db && uv run python init_db.py

• Claude Desktop not recognizing MCP tools? Ensure paths in JSON config are absolute and correct.

Resources

• GitHub Repository
• Canvas API Documentation
• MCP Documentation