Does onedrive_search_files search inside subfolders?

Yes. The search is recursive — it walks the directory you point it at and every subfolder beneath it, matching file names along the way and returning up to 50 results by default.

Do I need a Microsoft account or API key?

No. LMCP searches the OneDrive folder already synced on your Mac through the local filesystem. There are no Graph API tokens, OAuth flows, or cloud credentials to configure.

Is my file data sent to the cloud?

No. The search runs entirely on your Mac. File names and contents never leave your device and nothing is stored on an LMCP server, keeping you GDPR-compliant by design.

Search OneDrive Files by Name with AI on Mac

The onedrive_search_files tool lets your AI assistant search a OneDrive folder by file name and find matches anywhere inside it, recursively. It walks the directory tree, matches file names against your query, and returns up to a set number of results (50 by default). You simply describe what you are looking for in plain language and the AI does the digging. For example, you can type: "Search my OneDrive Documents folder for any file with 'invoice 2024' in the name and list where each one lives."

Because it runs through LMCP, a free, native macOS MCP server with 150+ tools, the search happens on your own Mac against your real OneDrive sync folder. There is no Microsoft Graph API token to set up, no cloud round-trip, and nothing copied off your machine. Download LMCP to get started.

Which AI agents work?

LMCP works with every major MCP-capable AI client:

Desktop clients (auto-configure via local stdio): Claude Desktop, Cursor, VS Code (GitHub Copilot), Windsurf and Zed.
Web AIs (connect through the LMCP Cloud Relay connector): ChatGPT, Claude.ai, Grok and Perplexity.

Desktop apps pick up the tool automatically once LMCP is installed. Web assistants link to your Mac securely through the Cloud Relay, so the same OneDrive search works from a browser tab.

Automation

File search is rarely the end goal — it is the first step in a chain. The AI can take the paths returned by onedrive_search_files and feed them straight into other LMCP tools: read a matched document, summarize a found PDF, attach the file to an email, move it to another folder, or list its metadata. A single prompt like "Find every contract in OneDrive named with 'NDA' and email me a summary of each" becomes search, then read, then compose — all without you opening a Finder window.

Context

The AI is working with your actual Mac data, not a generic web answer. It searches the real OneDrive folder synced on your machine, so results reflect exactly what you have right now: the latest files, the real folder structure, the true names. That grounding is what makes the assistant useful for your work instead of guessing — it sees your files the way you do.

Productivity

Hunting for a file by memory across nested OneDrive folders can eat several minutes each time, and Spotlight does not always cover synced cloud paths cleanly. With onedrive_search_files, you ask once and get every match with its full path in seconds. Multiply that across a day of "where did I put that report" moments and it adds up to real time back. It also shines when you only remember part of a name — the recursive match finds it wherever it ended up.

Privacy & GDPR

Everything runs locally on your Mac. The directory walk, the name matching, and the results all stay on your device — no file content or file names are ever sent to a server or stored in the cloud by LMCP. There are no API keys and no third-party accounts. This local-only design makes LMCP GDPR-compliant by architecture: data that never leaves the machine cannot be mishandled in transit or at rest elsewhere.