edstem-mcp
Exposes the full Ed Discussion API to MCP clients, enabling users to manage threads, comments, and course activity through natural language. It provides comprehensive tools for searching and posting content, featuring automatic conversion from markdown to Ed's internal XML format.
README
edstem-mcp
MCP server for Ed Discussion — expose Ed's full API to Claude and other MCP clients.
Setup
npm install
npm run build
Set your API token (get one at https://edstem.org/us/settings/api-tokens):
export ED_API_TOKEN=your_token
export ED_REGION=us # optional: us (default), au, etc.
Claude Code
Add to ~/.claude.json:
{
"mcpServers": {
"edstem": {
"command": "node",
"args": ["/path/to/edstem-mcp/dist/index.js"],
"env": {
"ED_API_TOKEN": "your_token"
}
}
}
}
Tools (22)
| Tool | Description |
|---|---|
get_user |
Get authenticated user info and enrolled courses |
list_threads |
List threads in a course (sortable, paginated) |
get_thread |
Get thread by global ID with comments |
get_course_thread |
Get thread by course-local number (the # in the UI) |
search_threads |
Search threads by title, content, or category |
post_thread |
Create a new thread (supports markdown input) |
edit_thread |
Edit an existing thread |
lock_thread |
Lock a thread |
unlock_thread |
Unlock a thread |
pin_thread |
Pin a thread |
unpin_thread |
Unpin a thread |
endorse_thread |
Endorse a thread |
unendorse_thread |
Remove thread endorsement |
star_thread |
Star/bookmark a thread |
unstar_thread |
Remove star |
post_comment |
Post a comment or answer on a thread |
reply_to_comment |
Reply to an existing comment |
endorse_comment |
Endorse a comment |
unendorse_comment |
Remove comment endorsement |
accept_answer |
Accept a comment as the answer |
list_users |
List course roster (staff/admin) |
list_user_activity |
List a user's threads and comments |
upload_file_from_url |
Upload a file to Ed from a URL |
format_content |
Preview markdown to Ed XML conversion |
Resources (2)
| Resource | URI | Description |
|---|---|---|
| User Info | edstem://user |
Authenticated user details |
| Courses | edstem://courses |
Enrolled courses list |
Prompts (3)
| Prompt | Description |
|---|---|
check_assignment |
Look up assignment details, requirements, and staff clarifications |
unanswered_questions |
List unresolved questions in a course |
my_activity |
Show your recent posts and comments in a course |
Content Format
Thread and comment content uses Ed's XML document format. This server auto-converts markdown to Ed XML, so you can write content naturally:
# Heading
**Bold** and *italic* text with `inline code`
- Bullet list
- Items
1. Numbered
2. List
> [!info] This becomes an Ed callout
Pass raw Ed XML (starting with <document) to bypass conversion.
Testing
npm test
Uses Node's built-in test runner (node:test). Tests cover the markdown-to-XML content converter and API client (URL construction, headers, error handling).
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。