canvas-mcp

This repository contains a Message Control Protocol (MCP) server implementation for interacting with the Canvas Learning Management System API. The server is designed to work with Claude Desktop and potentially other MCP clients.

The Canvas MCP Server provides a local interface to Canvas LMS API, allowing you to:

• List and manage courses
• Access assignments and submissions
• View announcements
• Retrieve course syllabi and modules
• Manage users and enrollments
• Generate course summaries

• Python 3.x
• Virtual environment (venv)
• Canvas API Token
• Canvas API URL (e.g., https://canvas.illinois.edu/api/v1)

1. Clone this repository:

git clone https://github.com/vishalsachdev/canvas-mcp.git
cd canvas-mcp

2. Create and activate a virtual environment:

python -m venv canvas-mcp
source canvas-mcp/bin/activate  # On Unix/macOS

3. Install dependencies:

pip install -r requirements.txt

Create a .env file in the root directory with the following variables:

CANVAS_API_TOKEN=your_canvas_api_token_here
CANVAS_API_URL=https://canvas.youruniversity.edu/api/v1

Replace the values with:

• Your Canvas API Token (How to get your Canvas API token)
• Your university's Canvas API URL

The start_canvas_server.sh script is already configured to:

• Load environment variables from the .env file
• Activate the virtual environment
• Run the cached server implementation

Make the start script executable:

chmod +x start_canvas_server.sh

1. Install Claude Desktop if you haven't already.

2. Create or edit the Claude Desktop configuration file:

vim ~/Library/Application\ Support/Claude/claude_desktop_config.json

3. Add the Canvas MCP server configuration:

{
  "mcpServers": [
    {
      "name": "canvas-api",
      "command": "/Users/YOUR_USERNAME/path/to/canvas-mcp/start_canvas_server.sh"
    }
  ]
}

Replace /Users/YOUR_USERNAME/path/to/canvas-mcp with the absolute path to where you cloned this repository.

4. Restart Claude Desktop to load the new configuration.

The server provides the following tools for Canvas LMS interaction:

• list_courses: List all courses for the authenticated user
• get_course_details: Get detailed information about a specific course
• summarize_course: Generate a comprehensive summary of a course

• list_assignments: List all assignments for a course
• get_assignment_details: Get detailed information about a specific assignment
• get_assignment_description: Get the full description of an assignment

• list_submissions: List all submissions for a specific assignment

• list_users: List all users enrolled in a course

• list_announcements: List all announcements for a course
• get_course_syllabus: Get the syllabus for a course
• get_course_modules: Get all modules for a course

This MCP server is designed to work seamlessly with Claude Desktop:

1. Claude Desktop will automatically start the server when needed
2. You'll see the Canvas API tools in the Claude Desktop interface (hammer icon 🔨)
3. You can ask Claude to perform Canvas operations like "Show me my courses" or "Get the syllabus for my Biology course"

For manual testing, you can start the server directly:

./start_canvas_server.sh

The server uses:

• fastmcp: A Python library for building MCP servers
• httpx: For asynchronous HTTP requests to the Canvas API
• Caching mechanisms to improve performance for course lookups

The main implementation file is canvas_server_cached.py, which provides:

• Efficient caching of course information
• Pagination handling for Canvas API requests
• Error handling and reporting
• Support for both course IDs and course codes

The server requires the following Python packages:

• httpx: For HTTP requests
• fastmcp: For MCP server implementation
• requests: For some HTTP operations
• Other standard libraries for encoding and networking

If you encounter issues:

1. Server Won't Start

   • Check that your .env file exists and contains valid credentials
   • Verify the virtual environment path in start_canvas_server.sh
   • Ensure all dependencies are installed

2. Authentication Errors

   • Verify your Canvas API token is valid and not expired
   • Check that you have the necessary permissions in Canvas

3. Connection Issues

   • Ensure your Canvas API URL is correct
   • Check your internet connection
   • Verify your institution hasn't restricted API access

4. Debugging

   • Check the server logs in the Claude Desktop console
   • Try running the server manually to see error output

• Your Canvas API token grants access to your Canvas account
• Never commit your .env file to version control
• Consider using a token with limited permissions if possible
• The server runs locally on your machine and doesn't expose your credentials externally

Contributions are welcome! Feel free to:

• Submit issues for bugs or feature requests
• Create pull requests with improvements
• Share your use cases and feedback

This project is licensed under the MIT License - see the LICENSE file for details.

Created by Vishal Sachdev