Canvas LMS MCP Server
Exposes Canvas LMS tools for courses, assignments, and quizzes via a remote MCP server running on Cloudflare Workers. Enables querying active courses and upcoming assignments/quizzes through natural language.
README
Canvas LMS MCP Server (Cloudflare Workers)
Remote MCP server that connects to Canvas LMS and exposes course/assignment/quiz tools over Cloudflare Workers.
What this server provides
The MCP agent is served at:
GET /mcp(MCP endpoint)
It currently implements these tools:
get_active_courses: Returns active courses filtered to course names containingSPorFL.get_assignments: Returns assignments due in the nextndays (1-90, default7).get_quizzes: Returns quizzes due in the nextndays (1-90, default7).
Prerequisites
- Node.js 18+ (or newer)
- npm
- A Cloudflare account
- A Canvas LMS access token and base URL (for example
https://sjsu.instructure.com)
Install
npm install
Configure environment variables
Use local vars for development by creating/updating .dev.vars:
CANVAS_BASE_URL=https://your-school.instructure.com
CANVAS_TOKEN=your_canvas_token
For deployment, use Wrangler secrets instead of committing credentials:
wrangler secret put CANVAS_TOKEN
wrangler secret put CANVAS_BASE_URL
Run locally
npm run dev
Default local URL:
http://localhost:8787/mcp
Health check route:
http://localhost:8787/
Deploy
npm run deploy
After deploy, your MCP endpoint is:
https://<your-worker>.<your-subdomain>.workers.dev/mcp
Use with an MCP client (example: Claude Desktop via mcp-remote)
Add/update your MCP client config:
{
"mcpServers": {
"canvas": {
"command": "npx",
"args": [
"mcp-remote",
"https://<your-worker>.<your-subdomain>.workers.dev/mcp"
]
}
}
}
For local testing:
{
"mcpServers": {
"canvas": {
"command": "npx",
"args": ["mcp-remote", "http://localhost:8787/mcp"]
}
}
}
Useful scripts
npm run dev: Start local Worker development server.npm run deploy: Deploy to Cloudflare Workers.npm run type-check: Run TypeScript type checking.npm run lint:fix: Auto-fix lint issues via oxlint.npm run format: Format repo with oxfmt.npm run cf-typegen: Regenerate Wrangler type bindings.
Project structure
src/index.ts: MCP server definition and tool implementations.src/lib/helper.ts: Canvas pagination (Linkheader parsing).public/: Static assets for Worker static asset binding.wrangler.jsonc: Cloudflare Worker configuration.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。