A minimal Canvas LMS MCP (Machine Conversation Protocol) server for easy access to education data through your Canvas LMS instance. This server provides a bridge between AI systems (like Cursor) and Canvas Learning Management System.
To install Canvas LMS Server for Claude Desktop automatically via Smithery:
npx -y @smithery/cli install @ahnopologetic/canvas-lms-mcp --client claude
uv package manager (recommended)The easiest way to install and run canvas-lms-mcp is using uvx:
uvx canvas-lms-mcp
This will run the server in an isolated environment without installing it permanently.
To install the tool permanently:
uv tool install canvas-lms-mcp
Clone the repository:
git clone https://github.com/yourusername/canvas-lms-mcp.git
cd canvas-lms-mcp
Install with uv:
# Install uv if you don't have it yet
curl -LsSf https://astral.sh/uv/install.sh | sh
# Create a virtual environment and install dependencies
uv venv
uv pip install -e .
Alternatively, use traditional methods:
python -m venv .venv
source .venv/bin/activate # On Windows: .venv\Scripts\activate
pip install -e .
Set the following environment variables:
export CANVAS_API_TOKEN="your_canvas_api_token"
export CANVAS_BASE_URL="https://your-institution.instructure.com" # Default: https://canvas.instructure.com
You can get your Canvas API token from your Canvas account settings.
Start the server with uv:
uv run src/canvas_lms_mcp/main.py
If installed with uvx tool:
canvas-lms-mcp
By default, the server runs on http://localhost:8000. You can use the FastMCP interface at http://localhost:8000/docs to interact with the API.
The server provides the following tools for interacting with Canvas LMS:
list_coursesList courses that the user is actively enrolled in.
Parameters:
page (optional, default=1): Page number (1-indexed)items_per_page (optional, default=10): Number of items per pageget_courseGet a single course by ID.
Parameters:
course_id (required): Course IDinclude (optional): List of additional data to includeget_course_syllabusGet a course’s syllabus.
Parameters:
course_id (required): Course IDget_course_modulesGet modules for a course.
Parameters:
course_id (required): Course IDinclude (optional): List of additional data to includelist_assignmentsList assignments for a course.
Parameters:
course_id (required): Course IDbucket (required): Filter assignments by (“past”, “overdue”, “undated”, “ungraded”, “unsubmitted”, “upcoming”, “future”)order_by (required): Field to order assignments by (“due_at”, “position”, “name”)page (optional, default=1): Page number (1-indexed)items_per_page (optional, default=10): Number of items per pageget_assignmentGet a single assignment by ID.
Parameters:
course_id (required): Course IDassignment_id (required): Assignment IDlist_quizzesList quizzes for a course.
Parameters:
course_id (required): Course IDinclude (optional): List of additional data to includepage (optional, default=1): Page number (1-indexed)items_per_page (optional, default=10): Number of items per pageget_quizGet a single quiz by ID.
Parameters:
course_id (required): Course IDquiz_id (required): Quiz IDlist_filesList files for a course or folder.
Parameters:
course_id (optional): Course IDfolder_id (optional): Folder IDinclude (optional): List of additional data to includepage (optional, default=1): Page number (1-indexed)items_per_page (optional, default=10): Number of items per pagelist_planner_itemsList planner items for the authenticated user.
Parameters:
start_date (required): Start date in ISO 8601 formatend_date (required): End date in ISO 8601 formatcontext_codes (optional): List of context codes (e.g., [“course_123”])page (optional, default=1): Page number (1-indexed)items_per_page (optional, default=10): Number of items per pageCursor is an AI-powered IDE that can interact with the Canvas LMS MCP server to provide education data directly within your development environment.
Install the Cursor IDE from https://cursor.sh/
Create a .cursor/mcp.json file in your project directory with the following content:
{
"mcpServers": {
"canvas": {
"command": "uvx",
"args": [
"canvas-lms-mcp"
],
"env": {
"CANVAS_API_TOKEN": "your_canvas_api_token",
"CANVAS_BASE_URL": "https://your-institution.instructure.com"
}
}
}
}
Replace:
your_canvas_api_token with your actual Canvas API tokenyour-institution.instructure.com with your Canvas institution URLRestart Cursor for the changes to take effect.
You can also integrate a time server for timezone-related queries by adding a “time” server to your mcp.json:
"time": {
"command": "uvx",
"args": [
"mcp-server-time",
"--local-timezone=America/New_York"
]
}
This allows you to use time-related functions with your Canvas data.
Once connected, you can ask Cursor AI about your Canvas data:
Example conversation:
YOU: What assignments do I have due soon?
CURSOR: I'll check your upcoming assignments.
Based on your Canvas data, here are your upcoming assignments:
- "Final Project" for CS101 due on December 10, 2023
- "Lab Report #5" for BIOL200 due on December 7, 2023
- "Research Paper" for ENGL301 due on December 15, 2023
For detailed development instructions, please see the DEVELOPMENT.md{:target=“_blank”} file.
This project is licensed under the MIT License - see the LICENSE{:target=“_blank”} file for details.
