mcp-canvas-lms

概述

这是一个功能完整的 MCP 服务器，提供对 Canvas LMS 功能的广泛访问。它实现了适当的模型上下文协议支持，包括 stdio 和 HTTP 传输选项。该服务器涵盖学生（作业提交、成绩查询）、教师（课程创建、评分）和管理员（用户管理、报告）的完整工作流程自动化。它包括强大的错误处理、自动分页，并通过 TypeScript 实现保持类型安全。

试试问 AI

装完之后，这里有 5 个你可以让 AI 做的事：

• canvas_health_check

  Check API connectivity to Canvas

  试试:Is the Canvas API working?试试:Check Canvas connection status
• canvas_list_courses

  List all available courses

  试试:Show me my courses试试:List all courses I'm enrolled in
• canvas_get_assignment

  Get detailed information about a specific assignment

  试试:Get details for assignment 5试试:Tell me about the Biology essay assignment
• canvas_submit_assignment

  Submit assignment work (text/URL/files)

  试试:Submit my English essay试试:Turn in my Math homework
• canvas_get_user_grades

  Get your overall grades across courses

  试试:Show me my current grades试试:What's my GPA this term?
• canvas_create_course

  Create a new course in an account

  试试:Create a new course called Advanced Biology试试:Set up a Computer Science 101 course
• canvas_enroll_user

  Enroll a student in a course

  试试:Enroll John Smith in Physics 101试试:Add student to my course roster
• canvas_get_account

  Get details of a Canvas account

  试试:Get information about my main account试试:Show account details for Engineering
• canvas_list_account_courses

  List all courses in an account

  试试:Show all courses in the Science account试试:List Computer Science courses
• canvas_create_user

  Create a new user in an account

  试试:Create a new student user Jane Doe试试:Add instructor account for Dr. Smith
• canvas_create_account_report

  Generate a report for an account

  试试:Generate enrollment report for Spring term试试:Create grade analytics report
• canvas_post_to_discussion

  Post a response to a discussion topic

  试试:Post to the Philosophy discussion about ethics试试:Reply to the assignment question

可对比工具

安装

{
  "mcpServers": {
    "canvas-mcp-server": {
      "command": "npx",
      "args": ["-y", "canvas-mcp-server"],
      "env": {
        "CANVAS_API_TOKEN": "your_token_here",
        "CANVAS_DOMAIN": "your_school.instructure.com"
      }
    }
  }
}

npm install -g canvas-mcp-server
export CANVAS_API_TOKEN="your_token_here"
export CANVAS_DOMAIN="your_school.instructure.com"
canvas-mcp-server

docker run -d \
  --name canvas-mcp \
  -p 3000:3000 \
  -e CANVAS_API_TOKEN="your_token" \
  -e CANVAS_DOMAIN="school.instructure.com" \
  ghcr.io/dmontgomery40/mcp-canvas-lms:latest

FAQ

如何获取我的 Canvas API 令牌？

登录 Canvas → 帐户 → 设置 → 滚动到"已批准的集成" → 点击"+ 新访问令牌" → 输入描述并复制生成的令牌。

我需要什么权限才能执行管理功能？

对于帐户级操作，您的 Canvas API 令牌必须由您机构的 Canvas 管理员授予管理权限。

mcp-canvas-lms 对比