Microsoft 365 File Search MCP Server (SharePoint & OneDrive)

What is Microsoft 365 File Search MCP Server (SharePoint & OneDrive)?

A Model Context Protocol (MCP) server that provides advanced file search capabilities within Microsoft 365, enabling efficient file discovery, metadata analysis, and integration with business workflows via SharePoint and OneDrive.

How to use Microsoft 365 File Search MCP Server (SharePoint & OneDrive)?

Register an application in the Azure Portal, add delegated or application permissions Sites.Read.All and Files.Read.All, then configure the server with CLIENT_ID, CLIENT_SECRET, TENANT_ID, and REGION environment variables. Integrate with Claude Desktop by adding an entry to claude_desktop_config.json using the uv command pointing to the server script.

Key features of Microsoft 365 File Search MCP Server (SharePoint & OneDrive)

• Searches files across SharePoint and OneDrive via search_m365_files
• Retrieves file content by drive ID and file ID
• Uses local caching to speed up repeat file access
• Cached files stored in ./src/mcp_m365_filesearch/.local/downloads
• Reduces redundant API calls and respects rate limits
• Supports both delegated and application permissions for Microsoft Graph

Use cases of Microsoft 365 File Search MCP Server (SharePoint & OneDrive)

• Search for documents across an entire Microsoft 365 tenant
• Retrieve and analyze file content within AI assistants
• Integrate file metadata into business workflows
• Quickly access frequently used files without repeated API calls

FAQ from Microsoft 365 File Search MCP Server (SharePoint & OneDrive)

What alternatives exist for file search in Microsoft 365?

This server provides direct MCP integration for AI assistants, bypassing manual searching and offering programmatic access via Microsoft Graph API.

What dependencies or runtime requirements are needed?

Python 3 and the uv package manager are required. An Azure-registered application with Sites.Read.All and Files.Read.All permissions is mandatory.

Where are cached files stored locally?

Cached files are stored in the ./src/mcp_m365_filesearch/.local/downloads directory relative to the project root.

What are the known limits or performance considerations?

The server uses local caching to improve response times and reduce API calls, but no specific rate limits or file size caps are mentioned in the README.

How does authentication and transport work?

Authentication uses OAuth 2.0 client credentials (client ID, client secret, tenant ID) for Microsoft Graph. The server communicates over standard MCP transport via local stdio.