Does whatsapp_send_message need the WhatsApp API or a business account?

No. LMCP uses the WhatsApp app already running on your Mac, so there are no API keys, tokens, or business accounts to set up. It works with your personal WhatsApp out of the box.

How does the AI know which chat to send to?

The chat_id always comes from a previous whatsapp_list_chats call. The assistant looks up your real chats first and never fabricates phone numbers or recipients, so messages reach the exact conversation you intend.

Is my WhatsApp data sent to the cloud?

No. Everything happens locally on your Mac. Messages are sent through your installed WhatsApp app and nothing is stored on an LMCP server, which is what makes LMCP GDPR-compliant by architecture.

Send WhatsApp Messages with AI on Mac

The whatsapp_send_message tool lets your AI assistant send a text message to any WhatsApp chat directly from your Mac. You just describe what you want in plain language and the assistant handles the rest. For example, you can type: "Send a WhatsApp to my brother saying I'll be 15 minutes late for dinner." The AI finds the chat, drafts the message, and sends it through your real WhatsApp — no copy-pasting, no switching windows.

Under the hood, the chat_id always comes from a previous whatsapp_list_chats call, so messages go to the exact conversation you mean. The AI never invents recipients or guesses phone numbers — it works only with the chats already on your Mac.

LMCP is a free, native macOS MCP server with 150+ tools. Download LMCP and connect your favorite AI in minutes.

Which AI agents work?

Almost every major AI assistant can use whatsapp_send_message through LMCP:

Desktop clients — Claude Desktop, Cursor, VS Code (GitHub Copilot), Windsurf and Zed auto-configure via local stdio. No tunnels, no extra setup.
Web AIs — ChatGPT, Claude.ai (web), Grok and Perplexity connect through the LMCP Cloud Relay connector, which securely bridges the web assistant to the LMCP server running on your Mac.

Once connected, the assistant sees all 150+ LMCP tools and can call WhatsApp alongside Mail, Calendar, Contacts and more.

Automation

The real power comes from chaining tools. A typical flow: the AI calls whatsapp_list_chats to find the right conversation, grabs the chat_id, then calls whatsapp_send_message to deliver your text. Beyond WhatsApp itself, you can combine it with other domains — pull a meeting time from Calendar and message a group the new schedule, read a contact from Contacts and ping them, or summarize an email and forward the gist over WhatsApp. The assistant orchestrates these steps so a single request becomes a complete workflow.

Context

Because LMCP runs natively on your Mac, the AI works with your real data — your actual WhatsApp chats, your contacts, your conversation history. There's no separate sandbox or mock account. When you say "message the family group," the assistant resolves it against the chats genuinely sitting in your WhatsApp, which makes its actions accurate and trustworthy.

Productivity

Sending a quick WhatsApp usually means unlocking your phone or alt-tabbing to the desktop app, finding the chat, and typing. With LMCP your AI does it inline while you stay in your work. Coordinating a group, confirming an appointment, or firing off a "running late" note drops from a minute of context-switching to a single sentence. Across a busy day of dozens of small messages, that adds up to real time back.

Privacy & GDPR

Everything runs locally on your Mac. LMCP reads and sends WhatsApp messages through the app already installed on your machine — no message is ever copied to an LMCP server, and there are no API keys to hand over. Your data stays under your control by design. Learn more about why LMCP is GDPR-compliant by architecture.