Microsoft Graph MCP Server
Provides read-only access to Microsoft 365 services including SharePoint, OneDrive, Outlook, Teams, and Calendar through the Microsoft Graph API, enabling users to search, browse, and retrieve content across their M365 suite.
README
Microsoft Graph MCP Server
Read-only access to Microsoft 365 suite via Model Context Protocol.
Architecture
The diagram shows the complete system architecture including Claude Code integration, MCP server components, Azure AD authentication, Microsoft Graph API gateway, and M365 services (SharePoint, OneDrive, Outlook, Teams, Calendar).
Features
- SharePoint: Search sites, list files, get content
- OneDrive: Browse user files, search
- Outlook: Search emails, read messages
- Teams: List teams/channels, read messages
- Calendar: View events, search calendar
Setup
📖 For detailed setup instructions, see SETUP.md
Quick Start
- Azure AD App Registration: Register app with Application permissions
- Install:
pip install -r requirements.txt - Configure: Add credentials to
~/.claude.json - Test:
msgraph.test_connection()
Required Azure AD Permissions
Register an Azure AD app with these Application permissions:
Sites.Read.All- SharePointFiles.Read.All- OneDrive/SharePoint filesMail.Read- OutlookChannelMessage.Read.All- Teams messagesChat.Read.All- Teams chatsCalendars.Read- CalendarUser.Read.All- User directoryGroup.Read.All- Teams/Groups
Grant admin consent for all permissions.
2. Install Dependencies
cd ~/.claude/mcp-servers/msgraph-mcp
pip install -r requirements.txt
3. Configure Claude
Add to ~/.claude.json:
{
"mcpServers": {
"msgraph-mcp": {
"type": "stdio",
"command": "python3",
"args": ["/Users/thianseongyee/.claude/mcp-servers/msgraph-mcp/server.py"],
"env": {
"AZURE_CLIENT_ID": "your-client-id",
"AZURE_TENANT_ID": "your-tenant-id",
"AZURE_CLIENT_SECRET": "your-client-secret"
}
}
}
}
Usage
Access via msgraph module in code execution:
# Test connection
result = msgraph.test_connection()
# Search SharePoint files
result = msgraph.search_files("formulation xlsx", limit=10)
# List recent emails
result = msgraph.list_recent_emails("user@company.com", limit=20)
# Get Teams channels
teams = msgraph.list_teams()
channels = msgraph.list_channels(teams['items'][0]['id'])
# View calendar
result = msgraph.list_events("user@company.com", days_ahead=7)
Available Functions
Connection
test_connection()- Test API connectivitylist_users(query, limit)- Search usersget_user(email)- User details
SharePoint
search_sites(query, limit)- Find siteslist_site_contents(site_id, path)- Browse folderssearch_files(query, site_id, limit)- Search filesget_file_content(site_id, item_id)- Read fileget_file_metadata(site_id, item_id)- File info
OneDrive
list_user_files(email, path)- Browse OneDrivesearch_user_files(email, query, limit)- Search files
Outlook
search_emails(query, email, limit)- Search maillist_recent_emails(email, limit, folder)- Recent mailget_email(email, message_id)- Full email
Teams
list_teams(limit)- List teamslist_channels(team_id)- Team channelsget_channel_messages(team_id, channel_id, limit)- Messagessearch_channel_messages(team_id, channel_id, query)- Search
Calendar
list_events(email, days_ahead)- Upcoming eventssearch_events(email, query, limit)- Search calendarget_event(email, event_id)- Event details
Security
- Read-only: No write permissions
- Sandboxed: Limited Python builtins
- App-only auth: No user credentials stored
- Token managed: Automatic refresh, never exposed
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。