
SharePoint MCP Server
A popular open-source MCP server that connects AI assistants like Claude to Microsoft SharePoint via the Microsoft Graph API, enabling natural language operations on sites, lists, documents, folders, file uploads, searches, and more.
Overview
SharePoint MCP Server is a Model Context Protocol (MCP) server that bridges modern AI assistants (such as Claude Desktop/Code, Cursor, and others) with Microsoft SharePoint. It uses the official Microsoft Graph API to allow natural language interaction with your organization's SharePoint sites, document libraries, lists, files, and folders.
Instead of manually navigating the SharePoint web interface or writing Graph API calls, you can ask your AI to "find the latest project proposal document", "upload this file to the team site", "create a new list for bug tracking", or "summarize documents in the marketing folder" — and the AI executes it securely.
Multiple community implementations exist (including popular repos like DEmodoriGatsuO/sharepoint-mcp and others), with Microsoft also providing official tenant-level MCP support for SharePoint and OneDrive (with some legacy servers deprecated in March 2026).
Key Features
- Microsoft Graph API Integration: Secure access to sites, lists, libraries, files, and folders.
- Natural Language Operations: Search documents, read/write list items, upload/download files, manage folders, and more.
- Authentication: Supports OAuth 2.0 with browser-based login flow and token management.
- Tool Exposure: Dozens of MCP tools for common SharePoint tasks (list management, file operations, site navigation, etc.).
- Read/Write Capabilities: Most implementations support both querying and modifying content (with optional read-only safety modes in some variants).
- Compatibility: Works with Claude Desktop, Claude Code, Cursor, VS Code, and any MCP-compliant client.
- Enterprise Ready: Designed for organizational use with proper permissions scoping.
Installation & Setup
- Clone a popular repository (e.g.,
git clone https://github.com/DEmodoriGatsuO/sharepoint-mcp.git). - Install dependencies (usually Node.js/TypeScript or Python-based).
- Configure environment variables for your SharePoint site URL and authentication settings.
- Run the server locally.
- Add the MCP server configuration to your AI client (e.g., in Claude Desktop's settings or via
server.json). - Authenticate via the OAuth flow (browser popup for secure login).
Many implementations provide one-command setup via npx or pip install. Detailed guides, including app registration in Azure AD for tenant-wide access, are available in the respective repositories.
Use Cases
- Document Search & Retrieval: Quickly locate and summarize files across large SharePoint libraries.
- Content Management: Upload, organize, rename, or move documents using conversational commands.
- List & Data Management: Create/update SharePoint lists, add items, generate views, or export data.
- Team Collaboration Automation: Generate reports, copy lists between sites, or update navigation.
- Knowledge Workflows: Ask AI to analyze project documents, extract insights, or prepare summaries without leaving your coding environment.
- Enterprise Automation: Integrate with other MCP servers for full Microsoft 365 workflows (e.g., combined with OneDrive, Teams, or Outlook MCP).
Technical Details
- Protocol: Full compliance with Model Context Protocol (MCP).
- Backend: Typically built with TypeScript/Node.js or Python, leveraging Microsoft Graph SDK.
- Security: OAuth 2.0 with delegated or application permissions; many support read-only modes for safety.
- Status: Actively maintained community projects; Microsoft offers official tenant-scoped MCP endpoints for SharePoint/OneDrive.
Note: Always review permissions carefully and follow least-privilege principles when granting access to organizational SharePoint data.
The SharePoint MCP Server significantly boosts productivity for teams heavily using Microsoft 365, turning AI assistants into powerful SharePoint co-pilots.
For setup instructions, tool references, and the latest updates, check popular repositories such as DEmodoriGatsuO/sharepoint-mcp or search for other community variants. Microsoft's official documentation also covers native SharePoint MCP integration for Agent 365 and Copilot Studio.