Canvas Connect MCP Server

A Model Context Protocol (MCP) server for managing Canvas LMS courses. This server enables you to manage assignments, students, announcements, and course modules through the Canvas API.

Features

Assignment Management

• List all assignments with details (due dates, points, publish status)
• Create new assignments with descriptions and due dates
• View assignment submissions
• Update student grades and add comments

Student Management

• List all enrolled students
• View individual student grades across all assignments
• Track student progress

Announcements

• List recent course announcements
• Create new announcements with rich text

Module & Content Management

• List all course modules
• View module contents and items
• Create new modules

Discussion & Grading Tools

• View discussion posts and replies with word counts
• Get grading rubrics for assignments
• AI writing detection to check for AI-generated student submissions

File Submission Grading

• AI-powered grading of Word (.docx), Excel (.xlsx), and PowerPoint (.pptx) file submissions
• Automatically fetches assignment instructions and rubric from Canvas
• Extracts document content and evaluates it against the rubric using Claude
• Grade a single student or all students in one pass
• Optionally posts grades and feedback directly back to Canvas

Setup

1. Install Dependencies

First, create a virtual environment and install the package:

python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -e .

2. Get Canvas API Credentials

1. Log into your Canvas LMS instance
2. Go to Account → Settings
3. Scroll down to "Approved Integrations"
4. Click "+ New Access Token"
5. Give it a purpose (e.g., "MCP Server") and click "Generate Token"
6. Copy the token immediately (you won't be able to see it again)

3. Find Your Course ID

Your course ID is in the URL when viewing your course:

https://your-institution.instructure.com/courses/[COURSE_ID]

4. Configure Environment Variables

Copy the example environment file and fill in your details:

cp .env.example .env

Edit .env with your information:

CANVAS_API_URL=https://your-institution.instructure.com
CANVAS_API_TOKEN=your_api_token_here
CANVAS_COURSE_ID=your_course_id_here
<<<<<<< HEAD
CANVAS_COURSE_NAME=Section 1  # Optional label, shown by list_courses
ANTHROPIC_API_KEY=your_anthropic_api_key_here  # Required for AI detection feature

# Optional: teaching a second course (e.g. another section)?
# CANVAS_COURSE_ID_2=your_second_course_id_here
# CANVAS_COURSE_NAME_2=Section 2

If you set CANVAS_COURSE_ID_2, every tool accepts an optional course_id argument (raw course ID, slot number 1/2, or the label you set) to target that course instead of the default. Run list_courses to see what's configured.

5. Get Anthropic API Key (Optional - for AI Detection)

======= ANTHROPIC_API_KEY=your_anthropic_api_key_here # Required for AI detection and file grading features

### 5. Get Anthropic API Key (Optional - for AI Detection & File Grading)
>>>>>>> 18da1ebec02f2a805e5d3ad0d959ca453cd6eb65

The AI writing detection and file submission grading features require an Anthropic API key:

1. Go to [console.anthropic.com](https://console.anthropic.com/)
2. Create an account or log in
3. Navigate to API Keys
4. Create a new API key
5. Add credits to your account (Plans & Billing)
6. Add the key to your `.env` file as `ANTHROPIC_API_KEY`

**Note**: These features use Claude to analyze student work. Each analysis costs a small amount of API credits.

### 6. Configure Claude Desktop

Add the server to your Claude Desktop configuration file:

**macOS**: `~/Library/Application Support/Claude/claude_desktop_config.json`
**Windows**: `%APPDATA%\Claude\claude_desktop_config.json`

```json
{
  "mcpServers": {
    "canvas-connect": {
      "command": "/Users/anthony/Documents/code/canvas_connect/venv/bin/python",
      "args": ["-m", "canvas_connect.server"],
      "env": {
        "CANVAS_API_URL": "https://your-institution.instructure.com",
        "CANVAS_API_TOKEN": "your_api_token_here",
        "CANVAS_COURSE_ID": "your_course_id_here",
        "CANVAS_COURSE_ID_2": "your_second_course_id_here"
      }
    }
  }
}

Important: Update the paths and credentials with your actual values.

7. Restart Claude Desktop

After saving the configuration, restart Claude Desktop to load the MCP server.

Available Tools

Every tool below also accepts an optional course_id argument for selecting between multiple configured courses (see "Configure Environment Variables" above). Omit it to use the default course.

Course Tools

list_courses

List the Canvas courses configured for this server, with their IDs and names.

Assignment Tools

list_assignments

List all assignments in the course with their details.

create_assignment

Create a new assignment.

• name (required): Assignment name
• points (required): Points possible
• description: Assignment description (HTML supported)
• due_at: Due date in ISO format (e.g., "2024-12-31T23:59:59")
• published: Whether to publish immediately (default: false)

get_assignment_submissions

Get all submissions for a specific assignment.

• assignment_id (required): Assignment ID

update_grade

Update a student's grade for an assignment.

• assignment_id (required): Assignment ID
• user_id (required): Student user ID
• grade (required): Grade (number or letter)
• comment: Optional comment

Student Tools

list_students

List all students enrolled in the course.

get_student_grades

Get a student's grades across all assignments.

• user_id (required): Student user ID

Announcement Tools

list_announcements

List recent announcements in the course.

• limit: Number of announcements to retrieve (default: 10)

create_announcement

Create a new announcement.

• title (required): Announcement title
• message (required): Announcement message (HTML supported)
• published: Whether to publish immediately (default: true)

Module Tools

list_modules

List all modules in the course.

get_module_items

Get all items in a specific module.

• module_id (required): Module ID

create_module

Create a new module.

• name (required): Module name
• published: Whether to publish immediately (default: false)

Discussion Tools

get_discussion_posts

Get all posts and replies for a discussion topic with word counts.

• topic_id (required): Discussion topic ID
• user_id: Optional filter for a specific student's posts

get_rubric

Get the grading rubric for a specific assignment.

• assignment_id (required): Assignment ID

AI Detection Tools

check_ai_writing

Analyze a student's discussion post to detect if it was likely written by AI such as ChatGPT or Claude.

• topic_id (required): Discussion topic ID
• user_id (required): Student user ID to check

Returns:

• Likelihood assessment (Low/Medium/High)
• Confidence level
• Key indicators found in the text
• Summary of the analysis

Note: This feature requires an Anthropic API key. See setup instructions above.

File Grading Tools

grade_file_submission

Download and AI-grade a single student's file submission (Word .docx, Excel .xlsx, or PowerPoint .pptx). Fetches the assignment instructions and rubric from Canvas, extracts the document content, and uses Claude to evaluate the work against the rubric.

• assignment_id (required): Assignment ID
• user_id (required): Student user ID
• auto_update_grade: If true, automatically post the AI-recommended grade and feedback to Canvas (default: false)

grade_all_file_submissions

Download and AI-grade every student's file submission for an assignment (Word, Excel, or PowerPoint). For each student who submitted a file, extracts content and grades it against the assignment rubric using Claude. Returns a full report.

• assignment_id (required): Assignment ID
• auto_update_grades: If true, automatically post AI-recommended grades and feedback to Canvas for all students (default: false)

Note: Grading a full class may take several minutes. Both tools require an Anthropic API key.

Example Usage

Once configured, you can ask Claude to help manage your course:

• "List all assignments in my course"
• "Create a new assignment called 'Lab 3: Python Basics' worth 50 points, due next Friday"
• "Show me all submissions for assignment 12345"
• "Update the grade for student 67890 on assignment 12345 to 85 with a comment 'Great work!'"
• "List all students in my course"
• "Show me the grades for student 67890"
• "Create an announcement titled 'Office Hours Change' with the message 'Office hours moved to Tuesday 2-4pm'"
• "List all modules"
• "Show me what's in module 5"
• "Create a new module called 'Week 3: Control Structures'"
• "Show me all the discussion posts for the Introductions discussion"
• "Get the rubric for assignment 1027200"
• "Check if student 127920 used AI to write their discussion post"
• "Grade student 67890's Word document submission for assignment 12345"
• "Grade all file submissions for assignment 12345 and show me the results"
• "Grade all file submissions for assignment 12345 and automatically post the grades to Canvas"

Security Notes

• Your Canvas API token has the same permissions as your Canvas account
• Keep your .env file secure and never commit it to version control
• The token is stored in Claude Desktop's configuration - ensure your computer is secure
• Consider creating a dedicated Canvas account with limited permissions for the MCP server

Troubleshooting

Server not showing up in Claude Desktop

• Check that the path to Python in the config is correct (use which python while your venv is activated)
• Verify your environment variables are set correctly
• Check Claude Desktop logs for errors

Canvas API errors

• Verify your API token is valid
• Check that your course ID is correct
• Ensure your Canvas account has sufficient permissions
• Some institutions may restrict API access

Connection issues

• Verify the Canvas API URL is correct (should end with .instructure.com)
• Check your internet connection
• Ensure Canvas is not undergoing maintenance

Development

To run the server directly for testing:

python -m canvas_connect.server

License

MIT License - feel free to modify and use for your courses.

Contributing

This is a personal project for managing Canvas LMS courses. Feel free to fork and adapt for your needs.