WhatsApp Message Sender MCP Tool

What is WhatsApp Message Sender MCP Tool?

This MCP server sends WhatsApp messages via the Meta (Facebook) WhatsApp Business API and provides a contacts resource. It is built for developers who want to integrate WhatsApp messaging into AI assistants or automated workflows using the Model Context Protocol.

How to use WhatsApp Message Sender MCP Tool?

Install dependencies with uv pip install -e . in the cloned repository. Run the server with uv run python main.py for standalone mode, or add the whatsapp entry to your Claude Local configuration under mcpServers, setting environment variables META_ACCESS_TOKEN and META_PHONE_NUMBER_ID. The tool exposes a send_whatsapp_message tool (parameters: phone_number, message) and a contacts resource at resource://contacts.

Key features of WhatsApp Message Sender MCP Tool

• Send WhatsApp messages via Meta Business API
• Supports international phone number format
• Provides a contacts resource listing names and numbers
• Runnable in standalone or Claude Local mode
• Configured with environment variables for credentials

Use cases of WhatsApp Message Sender MCP Tool

• Automating outbound WhatsApp notifications from an AI assistant
• Integrating WhatsApp messaging into Claude Local workflows
• Retrieving a contact list from the server resource

FAQ from WhatsApp Message Sender MCP Tool

What are the runtime requirements?

Python 3.12 or higher, uv package manager, and Meta WhatsApp Business API credentials (access token and phone number ID).

How do I configure the Meta API credentials?

Set the environment variables META_ACCESS_TOKEN and META_PHONE_NUMBER_ID in the Claude Local configuration or provide them in your environment.

Does this tool support receiving messages?

No. The README only describes sending messages. Receiving is shown in demo screenshots but is not documented as a server capability.

What transports does the server use?

The README does not specify transports; it runs as a Python process invoked via uv run and communicates via stdio for MCP.

How can I troubleshoot common errors?

Ensure the mcp package is installed (uv pip install mcp==1.6.0), check environment variables are correctly set, and try clearing uv cache (uv cache clean) if errors occur with uv run.