Canvas MCP Server

A Model Context Protocol (MCP) server that enables AI assistants like Claude to interact with Canvas LMS. This server provides tools for managing courses, announcements, rubrics, assignments, modules, pages, and student data through the Canvas API.

🚀 Now available as a Claude Desktop Extension for one-click installation!

Features

• List active courses and their details

• Post announcements to courses

• View course rubrics

• Get student enrollment information

• Access assignment details and submissions

• View student submission history and comments

• Analyze rubric statistics

• Manage course modules and module items

• Manage course pages, including content, revisions, and rollbacks

• Manage quizzes, including questions and question groups

Desktop Extension Benefits 🎯

The Canvas MCP server is now available as a Claude Desktop Extension for the ultimate user experience:

• 🎯 One-Click Installation: No terminal commands, no manual configuration

• 🔒 Secure Configuration: API tokens stored securely in your OS keychain

• 🔄 Automatic Updates: Get new features and fixes automatically

• 📦 Zero Dependencies: Everything bundled - no Node.js installation required

• 🌍 Cross-Platform: Works seamlessly on macOS and Windows

• 🎨 Native UI: Configure settings through Claude Desktop's beautiful interface

Perfect for educators who want powerful Canvas integration without the technical complexity!

Student Data Privacy & Anonymization 🔒

This Canvas MCP server includes privacy-first anonymization for student data. By default, all student names and emails are anonymized to protect privacy, but you can request real data when needed using natural language.

How It Works

Default Behavior (Anonymous):

• Student names become: Student 1, Student 2, etc.

• Student emails become: student1@example.com, student2@example.com

• Same student always gets the same pseudonym across all API calls

• Teacher/admin names are never anonymized (preserved in comments)

Natural Language Control: You can easily switch to real data by asking for it naturally:

❌ Anonymous (default):
"List all students in course 123"
→ Returns: Student 1, Student 2, student1@example.com

✅ Non-anonymous:
"List all students in course 123, but show their actual names and emails"
"Get assignment submissions with real student names"  
"Show me student data with actual identities"
→ Returns: John Smith, Jane Doe, john.smith@university.edu

Affected Tools

The following tools support anonymization control:

• list-students - Student enrollment data

• list-assignments - When including student submission data

• list-assignment-submissions - All submission data

• list-section-submissions - Section-filtered submissions

• list-rubric-assessments - Rubric assessment data

Privacy Features

🔒 Privacy by Default: All student data is anonymized unless explicitly requested otherwise
🗣️ Natural Language: Just ask for "actual names" when you need them
👨‍🏫 Teacher Protection: Teacher/admin names are never anonymized
🔄 Consistent Mapping: Same student gets same pseudonym across all calls
🎯 Selective: Only anonymizes student data, preserves all other information

Why Teachers/Admins Aren't Anonymized

The anonymization system specifically targets student privacy while preserving educational context:

🎓 Educational Rationale:

• Students are the protected population - They're the vulnerable party whose privacy needs protection

• Teacher feedback context matters - Knowing which instructor provided feedback is pedagogically valuable

• Staff accountability - Teachers and admins are professional staff, not students requiring privacy protection

📝 Practical Examples:

Assignment Comments (with anonymization enabled):
✅ "Excellent analysis of the data trends! - Prof. Johnson"
❌ "I found this section confusing - Student 1" 
❌ "Thanks for the feedback! - Student 2"

🔍 Technical Implementation:

• Comments include an author.role field ('student', 'teacher', 'admin', 'ta')

• Only authors with role === 'student' get anonymized

• Staff roles preserve real names for educational context

This design ensures student privacy while maintaining the educational value of knowing which instructor provided specific feedback. If you need full anonymization including staff, you can modify the anonymizer logic or request this as a feature enhancement.

Technical Implementation

Each tool that handles student data includes an anonymous parameter:

• Default: anonymous: true (privacy-first)

• Override: Claude automatically sets anonymous: false when you request real names

• Preserved: All functional IDs and non-personal data remain unchanged

Example of how Claude interprets your requests:

• "with their actual names" → anonymous: false

• "show real emails" → anonymous: false

• "anonymized" → anonymous: true (default anyway)

• "hide student identities" → anonymous: true

Prerequisites

• Node.js (v16 or higher)

• A Canvas API token

• Canvas instance URL (defaults to "https://fhict.instructure.com")

Installation

Option 1: Desktop Extension (Easiest) 🚀

NEW! One-click installation with Claude Desktop Extensions:

1. Download the latest extension: canvas-mcp-1.0.8.mcpb

2. Double-click the .mcpb file to open with Claude Desktop

3. Click "Install" - that's it!

4. Configure your Canvas API token and base URL through the Claude Desktop UI

Benefits:

• ✅ One-click installation - no terminal required

• ✅ Automatic updates when new versions are available

• ✅ Secure configuration - API tokens stored in OS keychain

• ✅ No dependencies - everything bundled in the extension

• ✅ Cross-platform - works on macOS and Windows

Option 2: NPM Package (Recommended)

The easiest way to use this MCP server is via npm:

npm install -g @r-huijts/canvas-mcp

Or use directly with npx (no installation required):

npx @r-huijts/canvas-mcp

Option 3: From Source

1. Clone this repository and install dependencies:

   git clone <repository-url>
   cd canvas-mcp
   npm install

2. Create a .env file in the project root:

   CANVAS_API_TOKEN=your_canvas_api_token_here
   CANVAS_BASE_URL=https://your-canvas-instance.instructure.com

3. Build the TypeScript project:

   npm run build

4. Start the server:

   npm start

Claude Desktop Integration

1. Open Claude Desktop's configuration file:

   MacOS:

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

   Windows:

   code %AppData%\Claude\claude_desktop_config.json

2. Add the Canvas MCP server configuration:

   For NPM Package Installation (Recommended):

   {
     "mcpServers": {
       "canvas": {
         "command": "npx",
         "args": [
           "-y",
           "@r-huijts/canvas-mcp"
         ],
         "env": {
           "CANVAS_API_TOKEN": "your_token_here",
           "CANVAS_BASE_URL": "https://your-canvas-instance.com"
         }
       }
     }
   }

   For Source Installation:

   {
     "mcpServers": {
       "canvas": {
         "command": "node",
         "args": [
           "/path/to/canvas-mcp/build/index.js"
         ],
         "env": {
           "CANVAS_API_TOKEN": "your_token_here",
           "CANVAS_BASE_URL": "https://your-canvas-instance.com"
         }
       }
     }
   }

3. Restart Claude Desktop to apply changes

Installation Notes

Benefits of NPM Package Installation:

• ✅ No need to clone/build source code

• ✅ Automatic updates available via npm

• ✅ Simpler configuration

• ✅ Works across different operating systems

• ✅ The -y flag automatically accepts package installation prompts

Troubleshooting NPX Issues:

• If you encounter permission issues, try running Claude Desktop as administrator (Windows) or with sudo (macOS)

• On Windows, ensure your PATH includes npm/npx executables

• For corporate networks, you may need to configure npm proxy settings

Available Tools

list-courses

Lists all active courses for the authenticated user

• No required parameters

• Returns course names, IDs, and term information

post-announcement

Posts an announcement to a specific course

• Required parameters:

  • courseId: string

  • title: string

  • message: string

list-rubrics

Lists all rubrics for a specific course

• Required parameters:

  • courseId: string

• Returns rubric titles, IDs, and descriptions

list-students

Gets a complete list of students enrolled in a course

• Required parameters:

  • courseId: string

• Optional parameters:

  • includeEmail: boolean (default: false)

  • anonymous: boolean (default: true) - Whether to anonymize student names/emails

• Returns student names, IDs, and optional email addresses

• Privacy: Student data is anonymized by default (use "with actual names" to override)

list-assignments

Gets all assignments in a course with submission status

• Required parameters:

  • courseId: string

• Optional parameters:

  • studentId: string

  • includeSubmissionHistory: boolean (default: false)

  • anonymous: boolean (default: true) - Whether to anonymize student data in submissions

• Returns assignment details and submission status

• Privacy: Student submission data is anonymized by default

list-assignment-submissions

Gets all student submissions for a specific assignment

• Required parameters:

  • courseId: string

  • assignmentId: string

• Optional parameters:

  • anonymous: boolean (default: true) - Whether to anonymize student names/emails

• Returns submission details, grades, and comments

• Privacy: Student data is anonymized by default

list-section-submissions

Gets all student submissions for a specific assignment filtered by section

• Required parameters:

  • courseId: string

  • assignmentId: string

  • sectionId: string

• Optional parameters:

  • includeComments: boolean (default: true)

  • anonymous: boolean (default: true) - Whether to anonymize student names/emails

• Returns submission details filtered by the specified section

• Privacy: Student data is anonymized by default

list-sections

Gets a list of all sections in a course

• Required parameters:

  • courseId: string

• Optional parameters:

  • includeStudentCount: boolean (default: false)

• Returns section details with optional student count

post-submission-comment

Posts a comment on a student's assignment submission

• Required parameters:

  • courseId: string

  • assignmentId: string

  • studentId: string

  • comment: string

• Returns confirmation of the posted comment

get-rubric-statistics

Gets statistics for rubric assessments on an assignment

• Required parameters:

  • courseId: string

  • assignmentId: string

• Optional parameters:

  • includePointDistribution: boolean (default: true)

• Returns detailed statistics about rubric assessment criteria

list-modules

Lists all modules in a course, optionally including inline items

• Required parameters:

  • courseId: string

• Optional parameters:

  • includeItems: boolean (default: false)

• Returns module names, IDs, positions, published state, and optionally item summaries

list-module-items

Lists all items in a specific module

• Required parameters:

  • courseId: string

  • moduleId: string

• Returns item type, title, ID, position, and published state

toggle-module-publish

Toggles the published/unpublished state of a module

• Required parameters:

  • courseId: string

  • moduleId: string

• Returns confirmation of the new published state

list-pages

Lists all pages in a course by URL slug

• Required parameters:

  • courseId: string

• Returns page titles, URL slugs, IDs, and published state

get-page-content

Gets the content (HTML/body) of a specific page by URL slug

• Required parameters:

  • courseId: string

  • pageUrl: string (the page's URL slug, e.g. 'syllabus')

• Returns page title, slug, ID, published state, updated date, and HTML body

update-page-content

Updates (or creates) a page by URL slug

• Required parameters:

  • courseId: string

  • pageUrl: string

• Optional parameters:

  • title: string

  • body: string (HTML)

  • editingRoles: string (comma-separated roles)

• Returns confirmation and updated page info

list-page-revisions

Lists all revisions for a page

• Required parameters:

  • courseId: string

  • pageUrl: string

• Returns revision IDs, timestamps, and editor info

revert-page-revision

Reverts a page to a previous revision

• Required parameters:

  • courseId: string

  • pageUrl: string

  • revisionId: string

• Returns confirmation and new page state

Quizzes

list-quizzes

Lists all quizzes in a course

• Required parameters:

  • courseId: string

• Returns a list of quizzes with their details

get-quiz

Fetches metadata for a single quiz

• Required parameters:

  • courseId: string

  • quizId: string

• Returns the full quiz object

create-quiz

Creates a new quiz in a course

• Required parameters:

  • courseId: string

  • title: string

• Optional parameters:

  • description: string

  • quiz_type: "practice_quiz" | "assignment" | "graded_survey" | "survey"

  • due_at: string (ISO 8601 format)

  • points_possible: number

  • published: boolean

• Returns the newly created quiz object

update-quiz

Updates an existing quiz

• Required parameters:

  • courseId: string

  • quizId: string

• Optional parameters: same as create-quiz

• Returns the updated quiz object

delete-quiz

Deletes a quiz

• Required parameters:

  • courseId: string

  • quizId: string

• Returns confirmation of deletion

list-quiz-questions

Lists all questions for a quiz

• Required parameters:

  • courseId: string

  • quizId: string

• Returns a list of question objects

get-quiz-question

Fetches a single quiz question

• Required parameters:

  • courseId: string

  • quizId: string

  • questionId: string

• Returns the full question object

create-quiz-question

Creates a new question for a quiz

• Required parameters:

  • courseId: string

  • quizId: string

  • question: object (containing question_text, question_type, points_possible, etc.)

• Returns the newly created question object

update-quiz-question

Updates a quiz question

• Required parameters:

  • courseId: string

  • quizId: string

  • questionId: string

• Optional parameters:

  • question: object (with fields to update)

• Returns the updated question object

delete-quiz-question

Deletes a quiz question

• Required parameters:

  • courseId: string

  • quizId: string

  • questionId: string

• Returns confirmation of deletion

list-quiz-question-groups

Lists all question groups for a quiz

• Required parameters:

  • courseId: string

  • quizId: string

• Returns a list of question group objects

get-quiz-question-group

Fetches a single quiz question group

• Required parameters:

  • courseId: string

  • quizId: string

  • groupId: string

• Returns the full question group object

create-quiz-question-group

Creates a new question group for a quiz

• Required parameters:

  • courseId: string

  • quizId: string

  • quizGroup: object (containing name, pick_count, question_points)

• Returns the newly created question group object

update-quiz-question-group

Updates a quiz question group

• Required parameters:

  • courseId: string

  • quizId: string

  • groupId: string

• Optional parameters:

  • quizGroup: object (with fields to update)

• Returns the updated question group object

delete-quiz-question-group

Deletes a quiz question group

• Required parameters:

  • courseId: string

  • quizId: string

  • groupId: string

• Returns confirmation of deletion

Available Prompts

analyze-rubric-statistics

Analyzes rubric statistics for formative assignments in a course and creates visualizations

• Required parameters:

  • courseName: string (The name of the course to analyze)

• Creates two comprehensive visualizations:

  1. Grouped stacked bar chart showing score distribution per criterion across all assignments

  2. Grouped bar chart showing average scores per criterion for all assignments

• Provides comparative analysis across assignments and criteria

• Includes progression analysis and trend identification

Troubleshooting

Common Issues

1. Server not appearing in Claude Desktop:

   • Verify configuration file syntax

   • Check file paths are absolute

   • Ensure Canvas API token is valid

   • Restart Claude Desktop

2. Connection errors:

   • Check Canvas API token permissions

   • Verify Canvas instance is accessible

   • Review Claude's MCP logs:

     # MacOS
     tail -f ~/Library/Logs/Claude/mcp*.log
     # Windows
     type %AppData%\Claude\Logs\mcp*.log

Debug Logging

The server logs errors to stderr. These can be viewed in Claude Desktop's logs or redirected when running manually:

node build/index.js 2> debug.log

Development

This MCP server uses the latest Model Context Protocol TypeScript SDK (v1.11.3+) with the new tool registration pattern. Each tool is registered with the server using the server.tool() method, which takes the following parameters:

1. Tool name (string)

2. Tool description (string)

3. Input schema (Zod schema or plain object)

4. Execute function (async function that implements the tool logic)

Here's an example of how a tool is registered:

server.tool(
  "list-courses",
  "List all courses for the authenticated user",
  {},
  async () => {
    // Tool implementation...
    return {
      content: [
        {
          type: "text",
          text: "Tool response..."
        }
      ]
    };
  }
);

Security Notes

1. API Token Security:

   • Never commit your Canvas API token to version control

   • Use environment variables or secure configuration

   • Regularly rotate your API tokens

2. Permissions:

   • Use tokens with minimum required permissions

   • Review Canvas API access logs periodically

License

MIT License

Copyright (c) 2024

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.