MCP Outlook
Enables AI assistants to manage Microsoft Outlook emails, calendar events, contacts, and folders via COM automation.
README
MCP Outlook
<img src="https://api.iconify.design/material-symbols/mail-outline.svg?color=%23004080" width="60"> <img src="https://api.iconify.design/material-symbols/calendar-month-outline.svg?color=%23004080" width="60"> <img src="https://api.iconify.design/material-symbols/perm-contact-calendar.svg?color=%23004080" width="60"> <img src="https://api.iconify.design/material-symbols/folder-outline.svg?color=%23004080" width="60"> <img src="https://api.iconify.design/material-symbols/settings-account-box.svg?color=%23004080" width="60">
A Model Context Protocol (MCP) server for Microsoft Outlook integration.
Version: 1.0.3 | Documentation: DOCUMENTATION.md | Contributing: CONTRIBUTING.md | Examples: EXAMPLES.md
Overview
This MCP server provides AI assistants with the ability to interact with Microsoft Outlook, including:
- <img src="https://api.iconify.design/material-symbols/mail-outline.svg?color=%23004080" width="20"> Email Management: Read, search, send, and draft emails with HTML support and Outlook signatures
- <img src="https://api.iconify.design/material-symbols/attach-file.svg?color=%23004080" width="20"> Attachment Management: List, download, and send email attachments with robust detection
- <img src="https://api.iconify.design/material-symbols/calendar-month-outline.svg?color=%23004080" width="20"> Calendar Management: View, create, and search calendar events
- <img src="https://api.iconify.design/material-symbols/perm-contact-calendar.svg?color=%23004080" width="20"> Contact Management: View, create, and search contacts
- <img src="https://api.iconify.design/material-symbols/folder-outline.svg?color=%23004080" width="20"> Folder Management: List folders, search in custom folders, and view Outlook rules
- <img src="https://api.iconify.design/material-symbols/settings-account-box.svg?color=%23004080" width="20"> Auto-Learning Style: Automatically learns your email formatting preferences (font, size, color) from sent emails
Features
Detailed documentation for each feature: DOCUMENTATION.md
<img src="https://api.iconify.design/material-symbols/mail-outline.svg?color=%23004080" width="30"> 5 Email Tools (Documentation)
get_inbox_emails- Retrieve emails from inbox with filtering optionsget_sent_emails- Retrieve sent emailssearch_emails- Search emails across folders by subject, body, or sendersend_email- Send emails with CC/BCC support, HTML content, and auto-learned stylecreate_draft_email- Create draft emails without sending, with HTML and auto-learned style
Auto-Learning Style: When
OUTLOOK_AUTO_LEARN_STYLE=true, the server dynamically learns your email formatting style (font-family, font-size, color) from your most recent sent email each time you send or create a draft. This learned style is automatically applied to plain-text emails. No caching - learning happens live on every send/draft operation.
<img src="https://api.iconify.design/material-symbols/attach-file.svg?color=%23004080" width="30"> 3 Attachment Tools (Documentation)
get_email_attachments- List all attachments from an email (with robust detection)download_email_attachment- Download a specific attachment to disksend_email_with_attachments- Send emails with file attachments
Robust Detection: Attachment detection intelligently filters out inline images, email signatures, and embedded items to show only real file attachments. Uses ContentID detection, file size filtering, and type checking.
<img src="https://api.iconify.design/material-symbols/calendar-month-outline.svg?color=%23004080" width="30"> 3 Calendar Tools (Documentation)
get_calendar_events- Get upcoming calendar eventscreate_calendar_event- Create new calendar events with attendeessearch_calendar_events- Search events by subject or location
<img src="https://api.iconify.design/material-symbols/perm-contact-calendar.svg?color=%23004080" width="30"> 3 Contact Tools (Documentation)
get_contacts- Retrieve contacts with optional name filteringcreate_contact- Create new contactssearch_contacts- Search contacts by name, email, or company
<img src="https://api.iconify.design/material-symbols/folder-outline.svg?color=%23004080" width="30"> 3 Folder Tools (Documentation)
list_outlook_folders- List all Outlook folders (ultra-fast, no item counts)search_emails_in_custom_folder- Search in specific custom folders with date filteringlist_outlook_rules- List all Outlook rules with conditions and actions
Performance Optimizations
This MCP has been heavily optimized for large mailboxes and to minimize Outlook freezing:
- Folder caching - 45x faster on repeated searches
- Date filtering - Search only recent emails (default: 2 days)
- Direct indexing - Faster iteration without
items.Count - Reduced limits - Prevents long freezes (max 50 emails)
- Smart defaults - Optimized for daily usage
- Silent logging - Minimal log output for cleaner integration
See DOCUMENTATION.md - Performances for detailed performance information.
Installation
Quick Start Guide: QUICK_START.md | Full Installation Guide: DOCUMENTATION.md - Installation
Prerequisites
- Windows OS (required for COM automation)
- Microsoft Outlook installed and configured
- Python 3.10+
Setup
-
Clone or download this repository
-
Install dependencies:
pip install -r requirements.txt
Or using the project file:
pip install -e .
-
Verify Outlook is running and configured with an account
-
Test the installation:
python tests/test_connection.py
Usage
Running the Server
Run the MCP server directly:
python src/outlook_mcp.py
Or using FastMCP's built-in CLI:
fastmcp run src/outlook_mcp.py
Configuration for Cursor/Claude Desktop
Add this configuration to your MCP settings file:
For Cursor (~/.cursor/mcp.json or workspace settings):
{
"mcpServers": {
"outlook": {
"command": "python",
"args": [
"C:/Users/YOUR_USERNAME/source/repos/MCP/src/outlook_mcp.py"
],
"env": {}
}
}
}
For Claude Desktop (%APPDATA%/Claude/claude_desktop_config.json on Windows):
{
"mcpServers": {
"outlook": {
"command": "python",
"args": [
"C:/Users/YOUR_USERNAME/source/repos/MCP/src/outlook_mcp.py"
]
}
}
}
Important: Replace YOUR_USERNAME with your actual Windows username.
Testing the Server
You can test the server using FastMCP's interactive mode:
fastmcp dev src/outlook_mcp.py
This will open an interactive prompt where you can test the tools.
Tool Examples
More Examples: EXAMPLES.md - Real-world use cases and workflows
Reading Emails
# Get last 10 unread emails
get_inbox_emails(limit=10, unread_only=True)
# Search for emails about "meeting"
search_emails(query="meeting", folder="inbox", limit=20)
Sending Emails
# Send a simple email
send_email(
to="colleague@company.com",
subject="Meeting Follow-up",
body="Hi, following up on our meeting...",
importance="high"
)
# Send email with HTML content and Outlook signature
send_email(
to="colleague@company.com",
subject="Project Update",
html_body="<h1>Update</h1><p>Here are the details...</p>",
signature_name="My Signature"
)
# Create a draft with multiple recipients and signature
create_draft_email(
to="team@company.com",
subject="Project Update",
body="Here's the latest update...",
cc="manager@company.com",
signature_name="My Signature"
)
Calendar Management
# Get next 7 days of events
get_calendar_events(days_ahead=7)
# Create a meeting
create_calendar_event(
subject="Team Standup",
start_time="2025-01-15 09:00",
end_time="2025-01-15 09:30",
location="Conference Room A",
required_attendees="team@company.com",
reminder_minutes=15
)
# Search for meetings
search_calendar_events(query="standup", days_range=30)
Contact Management
# Get all contacts
get_contacts(limit=50)
# Search for a contact
search_contacts(query="John Smith")
# Create a new contact
create_contact(
full_name="Jane Doe",
email="jane.doe@company.com",
company="Acme Corp",
job_title="Product Manager",
mobile_phone="+1-555-1234"
)
Security & Permissions
- This server requires access to your Outlook data
- It uses Windows COM automation (no credentials stored)
- All operations are performed with your current Outlook profile's permissions
- Make sure Outlook is running and configured before starting the server
Troubleshooting
"Unable to connect to Outlook"
- Ensure Microsoft Outlook is installed and running
- Verify Outlook is configured with at least one email account
- Try restarting Outlook
"ImportError: No module named 'win32com'"
- Install pywin32:
pip install pywin32 - After installation, run:
python Scripts/pywin32_postinstall.py -install(if needed)
Permission Errors
- Run your terminal/IDE as Administrator (may be required for COM automation)
- Check that Outlook is not blocked by security policies
Date Parsing Issues
- Use ISO format for dates:
2025-01-15 14:00 - Supported formats: "YYYY-MM-DD HH:MM", "tomorrow 2pm", "next Monday 10am"
Development
Project Structure
mcp-outlook/
├── src/
│ ├── __init__.py
│ └── outlook_mcp.py # Main MCP server
├── pyproject.toml # Project configuration
├── requirements.txt # Dependencies
├── .gitignore
└── README.md
Adding New Tools
To add a new tool, use the @mcp.tool() decorator:
@mcp.tool()
def my_new_tool(param1: str, param2: int = 10) -> str:
"""
Tool description.
Args:
param1: Description of param1
param2: Description of param2 (default: 10)
Returns:
JSON string with results
"""
# Implementation
return json.dumps({"success": True, "data": "..."})
Running Tests
pytest
Code Formatting
black src/
ruff check src/
Limitations
- Windows Only: Uses COM automation which is Windows-specific
- Outlook Required: Microsoft Outlook must be installed and running
- Single Account: Works with the default Outlook profile only
- Performance: Large mailboxes may have slower search performance
Roadmap
Recently Added
- [x] HTML email support
- [x] Outlook signature integration
- [x] Silent logging for cleaner integration
- [x] Outlook rules listing
- [x] Auto-learning email style from sent emails
- [x] Recursive folder search
Planned Features
- [ ] Attachment download/upload support
- [ ] Meeting response handling (accept/decline/tentative)
- [ ] Out-of-office settings (get/set/disable)
- [ ] Task management integration
- [ ] Folder management (create, move, delete)
- [ ] Advanced filtering (flags, categories, custom properties)
- [ ] Email rules creation and modification
- [ ] Cross-platform support (investigate MAPI alternatives)
License
MIT License - See LICENSE file for details.
Contributing
Contributions are welcome! Please read our Contributing Guide for details on:
- How to set up your development environment
- Code style guidelines (Black + Ruff)
- How to submit pull requests
- Roadmap and planned features
Quick Start for Contributors:
- Fork the repository
- Create a feature branch (
git checkout -b feature/amazing-feature) - Make your changes and test thoroughly
- Commit with conventional commits (
feat:,fix:,docs:, etc.) - Push and create a Pull Request
See CONTRIBUTING.md for detailed instructions.
Documentation
- README.md (this file) - Overview and quick start
- DOCUMENTATION.md - Complete technical documentation
- EXAMPLES.md - Real-world examples and use cases
- QUICK_START.md - 5-minute setup guide
- CONTRIBUTING.md - How to contribute
- CHANGELOG.md - Version history
Support
For issues or questions:
- Create an issue: GitHub Issues
- Check existing issues for similar problems
- Provide details: Windows version, Outlook version, Python version, error logs
Before creating an issue:
- Run
python tests/test_connection.pyand include the output - Check the Troubleshooting section
- Review existing issues
Acknowledgments
- Built with FastMCP
- Uses pywin32 for COM automation
- Inspired by the MCP Atlassian server architecture
Note: This tool accesses your local Outlook data. Ensure you follow your organization's security policies when handling email and calendar data.
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。