Do I need to give LMCP my Microsoft 365 password or API keys?

No. LMCP connects through your own authenticated Microsoft 365 session on your Mac. There are no API keys to manage and nothing is sent to an LMCP server — the search runs locally against your account.

Can I search contacts by company, not just by name?

Yes. m365_search_contacts matches by name, email address, or company, so you can ask things like 'find everyone at Contoso' and get the full list of matching contacts.

Which AI assistants can use this tool?

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

Search Microsoft 365 Contacts with AI on Mac

m365_search_contacts lets your AI assistant search your Microsoft 365 (Outlook / Exchange Online) address book by name, email, or company — without you opening Outlook on the web or scrolling through People. You just ask in plain language and the assistant returns the matching contact records: display name, email addresses, phone numbers, job title, and company.

A concrete example of what you can type:

"Search my Microsoft 365 contacts for everyone at Contoso and give me their emails."

The AI calls m365_search_contacts behind the scenes, queries your real M365 directory, and hands back the results inline — ready to use in a draft email, a meeting invite, or a spreadsheet.

Which AI agents work?

LMCP exposes 150+ tools to any MCP-capable AI client. This tool works with:

Desktop clients (auto-configure via local stdio): Claude Desktop, Cursor, VS Code (GitHub Copilot), Windsurf, and Zed. LMCP wires itself into their MCP config automatically — no manual JSON editing.
Web AIs (connect through the LMCP Cloud Relay connector): ChatGPT, Claude.ai, Grok, and Perplexity. These reach your Mac securely through the relay so the same local tools are available from the browser.

Automation

Contact search rarely happens in isolation. Because every tool runs through the same assistant, m365_search_contacts chains naturally with the rest of LMCP's 150+ tools:

Find a contact, then draft and send an email to them with the Mail or Outlook tools.
Pull a colleague's details and create a calendar invite with them as an attendee.
Look up a company's people and add a follow-up reminder or a task.
Match a name to an email, then search Teams or Messages for your last conversation with them.

One natural-language request can become a multi-step workflow the AI carries out end to end.

Context

The assistant is working with your real Microsoft 365 directory — the actual contacts in your work tenant, not a generic guess or a stale export. That means correct, current email addresses and company info, so the email it drafts or the invite it sends goes to the right person the first time. The AI sees the same data you'd see in Outlook, because it's reading the same source.

Productivity

Looking up a contact by hand means switching to Outlook web, typing into People, waiting, copying an address, and switching back. Multiply that by every name in a thread and it adds up fast. Asking your AI to "find all my contacts at Acme and CC them" collapses that into a single sentence — seconds instead of minutes, with no app-switching and no copy-paste errors. For anyone who lives in email and meetings, this is a daily time saver.

Privacy & GDPR

Everything runs locally on your Mac. LMCP is a native macOS MCP server: the contact search executes against your Microsoft 365 account through your own authenticated session, results stream straight to your AI client, and nothing is stored on any LMCP server. There's no cloud copy of your address book, no third-party database, no API keys to hand around. This is privacy by architecture — read more about GDPR-compliant by architecture.

Download LMCP — it's free and native for macOS.