# @likr/canvas-lms-mcp > Model Context Protocol (MCP) Server for Canvas LMS Latest version **2.0.0** (published 2026-07-09) · MIT license · 0 weekly downloads ## Install ```sh npm install @likr/canvas-lms-mcp pnpm add @likr/canvas-lms-mcp yarn add @likr/canvas-lms-mcp bun add @likr/canvas-lms-mcp ``` Provides the command `canvas-lms-mcp`. ## Health **Score 55/100 (C)** — status: active. Positive: no vulnerabilities; recently updated; high maintenance score. Warnings: low downloads; no types; no esm support. ## Facts | | | |---|---| | Version | 2.0.0 | | Published | 2026-07-09 | | First published | 2026-07-04 | | Weekly downloads | 0 | | License | MIT | | TypeScript types | none | | Module format | CommonJS | | Dependencies | 3 | | Unpacked size | 1.3 MB | | Known vulnerabilities | 0 | | Install scripts | no | | GitHub stars | 0 | | Author | Yosuke Onoue | | Maintainers | likr | | Keywords | mcp, modelcontextprotocol, canvas, canvas-lms, lms | ## Links - npm: https://www.npmjs.com/package/@likr/canvas-lms-mcp - Repository: https://github.com/likr/canvas-lms-agent - Homepage: https://github.com/likr/canvas-lms-agent#readme - Issues: https://github.com/likr/canvas-lms-agent/issues - npm.io page: https://npm.io/package/@likr/canvas-lms-mcp ## Dependencies (3) - [axios](https://npm.io/package/axios.md) ^1.7.9 - [dotenv](https://npm.io/package/dotenv.md) ^16.4.7 - [@modelcontextprotocol/sdk](https://npm.io/package/@modelcontextprotocol/sdk.md) ^1.0.1 ## Alternatives - [exif-parser](https://npm.io/package/exif-parser.md) — 3.8M weekly downloads - [vite-plugin-compression](https://npm.io/package/vite-plugin-compression.md) — 569.5K weekly downloads - [pica](https://npm.io/package/pica.md) — 442.4K weekly downloads - [@reportportal/client-javascript](https://npm.io/package/@reportportal/client-javascript.md) — 408.8K weekly downloads - [@tldraw/state](https://npm.io/package/@tldraw/state.md) — 316.0K weekly downloads ## Recent versions - 2.0.0 (latest) — 2026-07-09 - 1.1.0 — 2026-07-04 - 1.0.0 — 2026-07-04 ## README # Canvas LMS MCP Server A Model Context Protocol (MCP) server that connects Large Language Models (LLMs) to Canvas LMS APIs, enabling AI agents to manage courses, assignments, student submissions, discussions, announcements, and more. ## Features - **Course Management**: List courses, search users, inspect sections, and check course lists. - **Assignment Operations**: Retrieve assignments, create/update assignments, manage override dates, and assign peer reviews. - **Submissions & Grading**: Retrieve student submissions, submit files/grades, add comments, and handle role-adaptive workflows (student vs. educator). - **Discussions & Announcements**: Create, read, and moderate discussion topics and announcements. - **Calendar & Planning**: Fetch planner items, calendar events, and blackout dates. --- ## Prerequisites - **Node.js**: Version 18 or higher. - **Canvas API Token**: Generate a token from your Canvas instance under **Account** > **Settings** > **Approved Integrations** > **New Access Token**. - **Canvas Base URL**: The URL of your Canvas instance (e.g., `https://canvas.instructure.com` or `https://.instructure.com`). --- ## Configuration & Usage The easiest way to run the MCP server is via `npx`, which downloads and runs the server automatically. ### 1. Claude Desktop Add the following configuration 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-lms": { "command": "npx", "args": [ "-y", "@likr/canvas-lms-mcp" ], "env": { "CANVAS_BASE_URL": "https://canvas.instructure.com", "CANVAS_API_TOKEN": "YOUR_CANVAS_API_TOKEN_HERE" } } } } ``` ### 2. Cursor 1. Open Cursor and go to **Settings** > **Features** > **MCP**. 2. Click **+ Add New MCP Server**. 3. Configure the following: - **Name**: `canvas-lms` - **Type**: `command` - **Command**: `npx -y @likr/canvas-lms-mcp` 4. Set the following Environment Variables in the Cursor interface: - `CANVAS_BASE_URL`: `https://canvas.instructure.com` (or your custom instance URL) - `CANVAS_API_TOKEN`: `YOUR_CANVAS_API_TOKEN_HERE` ### 3. Cline / Roo-Code Add the server configuration to your Cline MCP settings file (typically at `~/.code/User/globalStorage/saoudrizwan.claude-dev/settings/cline_mcp_settings.json` or similar): ```json { "mcpServers": { "canvas-lms": { "command": "npx", "args": [ "-y", "@likr/canvas-lms-mcp" ], "env": { "CANVAS_BASE_URL": "https://canvas.instructure.com", "CANVAS_API_TOKEN": "YOUR_CANVAS_API_TOKEN_HERE" } } } } ``` --- ## Canvas LMS Skill This repository provides a dedicated `canvas-lms` skill for AI agents (supporting Antigravity or standard skill configurations). This skill provides the agent with structured best practices, role-adaptive behaviors (student vs. educator workflows), and reference guides to interact with the Canvas LMS APIs effectively. To install this skill in your agent workspace, run: ```bash npx skills add likr/canvas-lms-agent@canvas-lms ``` --- ## Environment Variables | Variable | Description | Default | | --- | --- | --- | | `CANVAS_BASE_URL` | The domain of your Canvas LMS instance (e.g. `https://canvas.instructure.com`). | `https://canvas.instructure.com` | | `CANVAS_API_TOKEN` | Your Canvas API personal access token. **(Required)** | None | --- ## Development If you wish to run the server locally for development: 1. Clone the repository and navigate to the directory: ```bash git clone https://github.com/likr/canvas-lms-agent.git cd canvas-lms-agent/canvas-lms-mcp ``` 2. Install dependencies: ```bash npm install ``` 3. Set up environment variables in a `.env` file: ```env CANVAS_BASE_URL=https://canvas.instructure.com CANVAS_API_TOKEN=your_token_here ``` 4. Start the server: ```bash npm start ``` 5. Run tests: ```bash npm test ``` ## License This project is licensed under the MIT License. See the [LICENSE](LICENSE) file for details. --- _Source: https://npm.io/package/@likr/canvas-lms-mcp · Machine-readable twin of the npm.io package page. Health data is recomputed on every publish._