canvas-lms

Canvas LMS Plugin for Claude Code & Cowork

A Claude plugin that provides full Canvas LMS integration — manage pages, assignments, modules, discussions, grades, rubrics, and more through natural language.

Prerequisites

Install the Canvas MCP server:

pip install canvas-mcp

Setup

1. Get a Canvas API Token

1. Log into your Canvas instance
2. Go to Account > Settings
3. Scroll to Approved Integrations and click + New Access Token
4. Give it a name (e.g., "Claude Code") and generate
5. Copy the token — you won't see it again

2. Set Environment Variables

export CANVAS_API_TOKEN="your-token-here"
export CANVAS_API_URL="https://your-institution.instructure.com/api/v1"

Or add them to your shell profile (~/.zshrc, ~/.bashrc).

3. Install the Plugin

Claude Code (CLI):

claude plugin install ./canvas-lms-plugin

Cowork: Upload the plugin directory through Customize > Browse plugins > Upload.

4. Verify Connection

Use the /canvas:status command to verify your connection:

/canvas:status

Or with a specific course:

/canvas:status 12345

What's Included

MCP Server (~60 Canvas API tools)

Full Canvas REST API coverage:

Category	Tools
Pages	Create, edit, get, list, delete pages
Assignments	Create, update, get details, list, grade
Modules	Create, update, list, add/remove items
Discussions	Create topics, post entries, list
Files	Upload, download, list course files
Rubrics	Create, update, associate with assignments
Grades	View submissions, bulk grade
Users	List students, groups, analytics
Messaging	Send conversations, announcements
Peer Review	Assign, track, analyze reviews

Command

• /canvas:status — Verify API connection and list accessible courses

Skill

• canvas-builder — Generate Canvas-ready HTML with proper structure, accessibility, and inline styling

Example Usage

"Create a new page in course 12345 called 'Week 3 Overview' with a header,
learning objectives, and a schedule table for Tuesday/Thursday classes"

"Create an assignment called 'Case Analysis 1' worth 20 points,
due March 1st at midnight Central time, accepting file uploads"

"List all modules in course 12345 and show me what's in Module 3"

"Upload slides.pdf to the Week 3 folder in course 12345"

Configuration

All configuration is via environment variables:

Variable	Required	Description
CANVAS_API_TOKEN	Yes	Your Canvas API access token
CANVAS_API_URL	Yes	Your institution's Canvas API URL (e.g., https://canvas.instructure.com/api/v1)
MCP_SERVER_NAME	No	Server display name (default: canvas-api)
ENABLE_DATA_ANONYMIZATION	No	Anonymize student data (default: true)
API_TIMEOUT	No	API request timeout in seconds (default: 30)

See the full env.template for all options.

Links

• Canvas MCP Server (GitHub)
• Canvas MCP on PyPI
• Canvas API Documentation

License

MIT