How does the AI get the email ID to read?

Usually it doesn't need one from you. The AI first lists or searches your M365 inbox with a companion tool, picks the relevant message, then passes that message's ID to m365_read_email automatically. You can also paste an ID directly if you have one.

Does m365_read_email send my email content to the cloud?

No. The tool reads the message locally on your Mac through your existing Microsoft 365 session. The full content is handed to your AI client, but LMCP never stores or copies it to any server.

Which AI assistants can use this tool?

Claude Desktop, Cursor, VS Code (GitHub Copilot), Windsurf and Zed connect locally via stdio. ChatGPT, Claude.ai, Grok and Perplexity connect through the LMCP Cloud Relay connector.

Read Microsoft 365 Email with AI on Mac

The m365_read_email tool lets your AI assistant open and read the full content of a specific Microsoft 365 (Outlook / Exchange Online) email by its ID — subject, sender, recipients, full body and metadata. Instead of switching to Outlook, copying text and pasting it into a chat, you just ask. A typical prompt looks like: "Read the M365 email with ID AAMkAGI2 and summarize what the client is asking for." The AI pulls the real message from your mailbox and works with it directly.

It's part of LMCP, a free, native macOS MCP server that connects AI assistants to your real Mac apps and accounts — locally. There are 150+ tools spanning Mail, Calendar, Contacts, Microsoft 365, Notes, Reminders, Messages and more. Download LMCP to get started.

Which AI agents work?

LMCP works with every major AI client. Desktop clients auto-configure through a local stdio connection — no manual JSON editing in most cases:

Claude Desktop — auto-configures via local stdio
Cursor — auto-configures via local stdio
VS Code (GitHub Copilot) — auto-configures via local stdio
Windsurf — auto-configures via local stdio
Zed — auto-configures via local stdio

Web-based AIs connect through the secure LMCP Cloud Relay connector:

ChatGPT — via Cloud Relay connector
Claude.ai (web) — via Cloud Relay connector
Grok — via Cloud Relay connector
Perplexity — via Cloud Relay connector

Automation

The real power comes from chaining m365_read_email with other tools. List your M365 inbox, let the AI pick the relevant message, read its full body, then act on it — all in one request:

Find a thread with m365_list_emails, read it in full, then draft a reply with m365_send_email.
Read a meeting request email, then create the event with the Calendar tools.
Read an email that mentions a deadline, then add a Reminder so nothing slips.
Read a message with an attachment reference and cross-check it against a file in Finder or OneDrive.

Context

Because LMCP runs on your machine, the AI works with your actual Microsoft 365 mailbox — not a stale export or a copy-paste snippet. It sees the genuine sender, the full body, the timestamps and the thread. That means summaries, replies and decisions are grounded in real data, so you get answers that reflect what's truly in your inbox right now.

Productivity

Reading and triaging email is one of the biggest time sinks of the workday. With m365_read_email, a long, dense message becomes a two-line summary in seconds. Ask the AI to extract action items, identify the actual question buried in a wall of text, or pull out a quoted figure — without ever leaving your chat window. For anyone processing dozens of Outlook emails a day, that's minutes saved per message and far less context-switching.

Privacy & GDPR

Everything happens locally on your Mac. m365_read_email reads the message through your existing Microsoft 365 session on the device — there are no third-party API keys to manage, and no email content is copied to or stored on any LMCP server. Your mailbox data never leaves your machine. This local-first design makes LMCP GDPR-compliant by architecture: there is no cloud datastore to breach, audit or subpoena.