home / mcp / canvas mcp server

Canvas MCP Server

Collection of Canvas LMS and Gradescope tools for the model context protocol. Query your courses, find resources, and chat with your upcoming assignments.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "aryankeluskar-canvas-mcp": {
      "url": "https://canvas-mcp--aryankeluskar.run.tools",
      "headers": {
        "SNITHERY_API_KEY": "YOUR_API_KEY"
      }
    }
  }
}

Canvas MCP equips your AI agents with the ability to interact with Canvas LMS and Gradescope. It enables natural language queries to locate resources, fetch upcoming and graded items, and retrieve course and assignment data from Gradescope, making your workflows faster and more integrated.

How to use

You will connect an MCP client to one or more MCP endpoints to start using Canvas MCP. You can operate from a local stdio configuration using Smithery’s CLI, or you can connect via an HTTP endpoint exposed for your agent interface. Once connected, you can ask for relevant resources, query upcoming assignments with course breakdowns, and retrieve Gradescope data such as courses, modules, items, files, calendar events, and assignments. Use natural language prompts to request the data you need, and the MCP will return structured results you can act on.

How to install

Prerequisites you need before installation:

Node.js and npm or a compatible Node runtime
Access to Smithery CLI when using the preferred installation path

Choose an installation approach that matches your workflow. You can connect via an HTTP endpoint for ChatGPT or use a local SMP configuration with Smithery to run the MCP locally.

# Option A: Connect via HTTP endpoint (ChatGPT integration)
# You will use an HTTP URL to connect your agent to Canvas MCP
# The endpoint URL is provided in the integration steps

Manual configuration (local instance)

If you are setting up Canvas MCP on a local machine, create a configuration file and an environment file as shown below.

{
  "mcpServers": {
      "canvas": {
          "command": "npx",
          "args": [
              "-y",
              "@smithery/cli",
              "run",
              "@aryankeluskar/canvas-mcp"
          ]
      }
  }
}

Available tools

Find relevant resources

Search and retrieve resources relevant to a given natural language query, helping you discover documents, guides, and files within Canvas LMS and related resources.

Query upcoming assignments

Fetch upcoming assignments for a course and present a breakdown of due dates, statuses, and related details.

Get courses (Gradescope)

Query Gradescope to retrieve your courses and their associated data.

Get modules (Gradescope)

Fetch modules for a Gradescope course to understand structure and content.

Get module items (Gradescope)

Retrieve items within a Gradescope module for quick access to materials.

Get file url (Gradescope)

Obtain direct URLs for files stored in Gradescope for quick access.

Get calendar events (Gradescope)

Pull calendar events associated with Gradescope courses and assignments.

Get assignments (Gradescope)

List assignments within Gradescope courses for easy monitoring.