MCP Server - Microsoft 365 File Search (SharePoint & OneDrive)

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

An MCP server that provides advanced file search capabilities within Microsoft 365, enabling file discovery, metadata analysis, and integration with business workflows by making available content from SharePoint and OneDrive.

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

Configure it in Claude Desktop by adding an entry to claude_desktop_config.json using the uv command to run server.py. Required environment variables: CLIENT_ID, CLIENT_SECRET, TENANT_ID, and REGION. You must also register an application in Azure Portal with delegated or application permissions Sites.Read.All and Files.Read.All, then grant admin consent.

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

• Two tools: search_m365_files and get_file_content
• Local caching of downloaded files for faster repeat access
• Cache stored in ./src/mcp_m365_filesearch/.local/downloads
• Returns file metadata including content summary, drive ID, and file ID
• Supports binary stream content retrieval

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

• Search for files in Microsoft 365 by keyword or criteria
• Retrieve file content for analysis or processing
• Integrate SharePoint/OneDrive file discovery into AI assistants
• Reduce API calls with local caching for frequently accessed files

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

What Azure API permissions are required?

The app registration needs Sites.Read.All and Files.Read.All (delegated or application), with admin consent granted.

Where are cached files stored?

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

What inputs does get_file_content require?

It requires a driveid (string, unique parent drive identifier) and a fileid (string, unique file identifier).

How does the cache work?

When a file is requested via get_file_content, the server first checks the local cache. If found, it returns the file from disk without making a new API call.

What runtime is needed to run the server?

The server is run using uv and Python; the command references server.py in the source directory.