m365-copilot-mcp
Integrates Microsoft 365 enterprise data into GitHub Copilot and Claude Desktop using Microsoft 365 Copilot APIs. It enables natural language queries across SharePoint, OneDrive, emails, calendars, and Teams meetings while maintaining full enterprise permission enforcement.
README
m365-copilot-mcp
MCP server for Microsoft 365 Copilot APIs — bring enterprise data into GitHub Copilot & Claude Desktop.
What This Does
Query SharePoint, OneDrive, email, calendar, and Teams meetings through natural language—all with full Microsoft 365 permission enforcement.
| Tool | API | What It Does | Best For |
|---|---|---|---|
m365_retrieve |
Retrieval | Returns text chunks for YOUR AI to reason over | Custom RAG, deep analysis |
m365_chat |
Chat | M365 Copilot synthesizes an answer | Quick Q&A, people/calendar |
m365_meetings |
Meeting Insights | AI summaries, action items, mentions | Post-meeting follow-up |
m365_search |
Search | Semantic document discovery | Finding files |
m365_chat_with_files |
Chat + Files | Ask questions about specific documents | Summarizing known files |
Requirements
- Microsoft 365 Copilot license (required for API access)
- Python 3.11+
- Azure AD app registration with delegated permissions
Quick Start
1. Create Azure AD App Registration
- Go to Azure Portal → Azure Active Directory → App registrations
- Click "New registration"
- Name:
m365-copilot-mcp - Supported account types: Single tenant (or Multitenant)
- Redirect URI:
http://localhost:8400(Mobile and desktop applications) - Add API permissions (Delegated):
Sites.Read.AllMail.ReadPeople.Read.AllOnlineMeetingTranscript.Read.AllChat.ReadChannelMessage.Read.AllExternalItem.Read.AllFiles.Read.AllOnlineMeeting.Read
- Grant admin consent
2. Install
# Clone the repo
git clone https://github.com/renepajta/m365-copilot-mcp.git
cd m365-copilot-mcp
# Create virtual environment
python3 -m venv .venv
source .venv/bin/activate
# Install
pip install -e ".[dev]"
3. Configure
Create a .env file:
AZURE_CLIENT_ID=your-app-client-id
AZURE_TENANT_ID=your-tenant-id
4. Test Locally
# Run in HTTP mode for debugging
m365-copilot --http --port 8000
# Check health
curl http://localhost:8000/health
MCP Client Configuration
VS Code / GitHub Copilot
Add to your .vscode/mcp.json:
{
"servers": {
"m365-copilot": {
"type": "stdio",
"command": "uvx",
"args": ["--from", "git+https://github.com/renepajta/m365-copilot-mcp.git", "m365-copilot"],
"env": {
"PYTHONUNBUFFERED": "1",
"AZURE_CLIENT_ID": "your-app-client-id",
"AZURE_TENANT_ID": "your-tenant-id"
}
}
}
}
Claude Desktop
Add to claude_desktop_config.json:
{
"mcpServers": {
"m365-copilot": {
"command": "uvx",
"args": ["--from", "git+https://github.com/renepajta/m365-copilot-mcp.git", "m365-copilot"],
"env": {
"AZURE_CLIENT_ID": "your-app-client-id",
"AZURE_TENANT_ID": "your-tenant-id"
}
}
}
}
Example Queries
# Quick Q&A (m365_chat)
"Who owns the budget approval process?"
"Summarize emails from Contoso this week"
"What meetings do I have tomorrow?"
# Deep research (m365_retrieve)
"Find all ADRs related to microservices architecture"
"Get policy documents about data retention from 2024"
# Meeting follow-up (m365_meetings)
"What action items came out of the security review?"
"When was I mentioned in yesterday's standup?"
# Document discovery (m365_search)
"Find contracts mentioning liability caps"
"Q3 board presentation slides"
# File analysis (m365_chat_with_files)
"Summarize key risks in this proposal"
"Compare revenue projections between these two reports"
Authentication Flow
On first run, you'll be prompted to authenticate:
- Browser auth (default): Opens browser for Microsoft sign-in
- Device code (fallback): For headless environments, displays a code to enter at microsoft.com/devicelogin
Tokens are cached locally in ~/.m365-copilot-mcp/ for subsequent runs.
Environment Variables
| Variable | Required | Description |
|---|---|---|
AZURE_CLIENT_ID |
Yes | App registration client ID |
AZURE_TENANT_ID |
Yes | Azure AD tenant ID |
AZURE_CLIENT_SECRET |
No | Only for confidential clients |
M365_COPILOT_TIMEOUT |
No | Request timeout in seconds (default: 60) |
M365_COPILOT_CACHE_DIR |
No | Token cache location |
Troubleshooting
"Insufficient permissions"
- Ensure admin consent is granted for all API permissions
- Verify user has Microsoft 365 Copilot license
"Token expired"
- Delete
~/.m365-copilot-mcp/and re-authenticate
"Gateway timeout" on long queries
- Break complex queries into smaller parts
- Use
m365_retrievefor better control over scope
API Rate Limits
| API | Rate Limit | Notes |
|---|---|---|
| Chat API | Not documented | Subject to Graph throttling |
| Retrieval API | 200 req/user/hour | Max 25 results per request |
| Search API | 200 req/user/hour | Max 100 results per request |
| Meeting Insights | Standard Graph limits | Available ~4 hours post-meeting |
Development
# Run tests
pytest tests/ -v
# Lint
ruff check src/
# Type check
mypy src/
License
MIT
References
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.