
Microsoft Teams
io.github.SurgeEnterpriseAI由 GitHub 账号 SurgeEnterpriseAI 发布v1.0.0更新于 Oct 24, 2026
Connect Claude to Microsoft Teams — read/send messages, list channels, search chats.
发布方自己的服务:Microsoft Learn MCP
概览
将 Claude 连接到 Microsoft Teams,让助手列出团队和频道、读取和发送消息、搜索聊天并查看个人资料。
- 可以做什么
- 列出你所在的团队及团队中的频道,读取最近的频道消息和话题回复,并在话题中回复或发布新的频道消息。列出你的私聊和群聊,读取聊天消息,并发送私聊或群聊消息。在所有 Teams 中搜索消息,并获取你的 Microsoft 365 个人资料。
- 适用场景
- 当你希望 Claude 直接在 Microsoft Teams 中工作时适用:跟进频道或聊天中的对话、搜索历史消息,或在不切换应用的情况下起草并发送回复。
- 运行要求
- 以本地 Node.js 进程(npm 包)通过 stdio 运行,或作为托管的 HTTP/SSE 服务器运行。登录使用你自己的 Microsoft 365 工作账户,通过 OAuth 完成。自行托管需要注册 Azure AD 应用,配置客户端密钥和委托的 Microsoft Graph 权限,并通过 TEAMS_CLIENT_ID 和 TEAMS_CLIENT_SECRET 设置。
安装
在 SourceWeft 中
- 打开 控制台中的 Microsoft Teams,将其添加到工作区。
- 为需要使用其工具的对话启用该服务。
Desktop only,通过 STDIO。 STDIO 服务会启动本地进程,因此需要 SourceWeft 桌面宿主。
其他 MCP 客户端
参照 仓库 中的启动说明。
README
Teams MCP Server
Let Claude read and send your Microsoft Teams messages. An open-source MCP server that connects Claude Code, Claude Desktop and Claude Cowork to Microsoft Teams: list your teams and channels, read channel and chat messages, search across Teams, and post or reply — all by asking Claude in plain English.
- What it does: gives Claude 11 Teams tools through Microsoft Graph, using your own Microsoft 365 sign-in. Claude sees only what you can already see in Teams. Messages pass through; nothing is stored except your sign-in token on your machine.
- Who it's for: people who live in Teams and use Claude — developers, team leads, support and ops people who want "summarise #standup since Monday" or "reply to Priya's question in #release" without switching windows. Also teams that want to host one shared Teams connector for everyone in their company.
- Cost: free and MIT-licensed. You need a Microsoft 365 work or school account and permission to register an app in Azure (or an admin who does it for you once).
[Teams MCP Server — 76-second walkthrough]
▶ Watch the 76-second walkthrough: Teams MCP Server — how it works (captions on; transcript)
Install in 3 steps (Claude Code, on your own computer)
- Register an app in Azure (5 minutes, once). Azure Portal → App registrations → New registration → any name, Accounts in any organizational directory, redirect URI type Web =
http://localhost:3333/callback. Then add a client secret and the delegated Microsoft Graph permissions listed in Azure AD App Registration below. Copy the client ID and secret. - Add the server to Claude:
- Sign in once. Start Claude and ask "List my Teams." A browser window opens; sign in with your Microsoft work account. Done — the token is cached, so you won't be asked again.
Using Claude Desktop instead? Put the same command and env in claude_desktop_config.json (see Claude Desktop). Want one server for your whole company? See Self-host.
Using it (plain English)
You don't call the tools yourself — you just ask Claude, and Claude picks the right one. Things you can say:
Good to know:
- Your access, nothing more. Claude uses your sign-in, so it can only see teams, channels and chats you're in. Each person who connects signs in with their own account.
- You approve what gets sent. Claude asks permission before each tool call unless you've allowed that tool. A good setup: allow the read and search tools, and keep the three send/reply tools on "ask" so you see every message before it posts.
- Sign-in lasts. The token is saved in
.data/tokens(local) and refreshed silently. Delete that folder to sign out. - Search needs a work account. Microsoft search across Teams doesn't work for personal Microsoft accounts.
- If something fails, ask Claude "what was the error?" — the server returns Microsoft's own message (for example, a missing permission or admin consent).
How It Works
The user connects their own Microsoft 365 account. They see their own teams, their own channels, their own messages. Each user's tokens are isolated. The server is the bridge — it never stores messages, just passes them through.
Features
- One-click connect — visit
/connect, sign in with Microsoft, done - Multi-tenant — any Microsoft 365 organization, any user
- 11 tools — teams, channels, messages, chats, search, profile
- Dual transport — STDIO for Claude Code, HTTP/SSE for Cowork & remote
- Multi-user — each session gets isolated OAuth tokens
- Auto-detect — picks the right transport based on launch method
- Token caching — silent refresh, no re-login every session
- Retry logic — automatic retries on throttled Graph API calls
- TypeScript — fully typed, clean architecture, MIT licensed
For Users: Connect in 1 Click
If someone is already hosting this server for your company:
- Claude asks you to connect Microsoft Teams
- You click the auth link
- Sign in with your Microsoft work account
- Done — Claude can now read your Teams
No API keys, no Azure portal, no setup.
For Developers: Self-Host
1. Clone & Install
2. Azure AD App Registration
- Go to Azure Portal → App registrations → New registration
- Name:
Teams MCP Server - Supported account types: "Accounts in any organizational directory (Any Azure AD directory - Multitenant)"
- Redirect URI: Web →
https://your-server.com/auth/callback(hosted) and/orhttp://localhost:3333/callback(local STDIO sign-in) - Click Register
Configure:
- Certificates & secrets → New client secret → copy the Value
- API Permissions → Add Delegated (Microsoft Graph):
ChannelMessage.Read.All,ChannelMessage.SendChat.Read,Chat.ReadWrite,ChatMessage.SendTeam.ReadBasic.All,Channel.ReadBasic.AllUser.Read
- Click Grant admin consent (for your own tenant; other tenants consent on first login)
3. Configure & Deploy
Deploy to Railway, Render, Azure App Service, or any Node.js host:
4. Submit to MCP Registry
Once deployed and tested, submit your server URL to Anthropic's MCP connector registry. Users will then see "Microsoft Teams" as a connect option in Claude Cowork — just like Slack and Gmail.
For Claude Code Users: Local STDIO Mode
Quickest: claude mcp add as in Install in 3 steps. Or add to your .mcp.json from a clone:
First run will open your browser for Microsoft login. After that, tokens are cached.
Claude Desktop
In claude_desktop_config.json:
Available Tools
Architecture
Environment Variables
For AI assistants
A short, structured summary for LLMs and AI search lives in docs/llms.md.
License
MIT — LICENSE
Contributing
Issues and PRs welcome. Built by SurgeEnterpriseAI — Surge Software Solutions, Bengaluru.
来源:README.md,提交 a6475b5
工具
0版本历史
1- v1.0.0最新Sep 16, 2026
评分
已满足适用检查项的 50%
该评分衡量此收录条目的完整程度,依据注册中心条目、端点与公开来源计算得出。它不是对服务器本身的评价,也不表示服务器安全或当前可用:可连通性是带日期的检查结果,评分也不包含安装量或用户评价。
- 近 30 天内可连通25 分
请保持端点能够响应 MCP 握手;SourceWeft 每 30 天探测一次。
- 已声明认证方式10 分
请在注册中心条目中声明凭据请求头,或保持端点能够响应探测,以便判定其认证方式。
- 已列出工具15 分
请让服务器在无凭据时响应 tools/list,或发布可在沙箱中无凭据启动的软件包。
- 描述必需15 分
- 图标10 分
- 文档15 分
- 许可证10 分
其他 Microsoft 集成

Microsoft Learn MCP
com.microsoft官方发布者
一个免费的远程 MCP 服务器,让 AI 助手搜索微软官方 Learn 文档、查找代码示例,并把文档页面抓取为 markdown。

Playwright Cloud Browsers
com.microsoft官方发布者
远程 MCP 服务器,为 AI 智能体提供托管的 Playwright 云端浏览器,可创建会话、访问并检查页面、与页面元素交互。

Microsoft Sentinel Data Exploration
com.microsoft官方发布者
远程 Microsoft Sentinel MCP 服务器,让助手用自然语言查找相关表并从 Sentinel 数据湖检索安全数据。
Visio
com.microsoft
让助手通过桌面自动化在用户的 Windows 机器上创建、更新和读取 Microsoft Visio 图表。

Microsoft Clarity MCP
由 GitHub 账号 surendranb 发布
本地 MCP 服务器,把 Microsoft Clarity 的行为 UX 数据(暴怒点击、无效点击、快速返回、JS 错误)提供给 AI 助手分析。