Meta launches WhatsApp Business MCP for AI powered developer workflows

MUMBAI: WhatsApp is getting a little more hands-on with AI, as Meta gives coding agents the keys to its business messaging toolbox. Meta has introduced WhatsApp Business Tools MCP, a new Model Context Protocol (MCP) server that allows AI coding agents to connect directly with the WhatsApp Business Platform and handle several parts of account setup, configuration and testing.

The idea is to take developers out of the usual tab-hopping routine. Instead of moving between Meta’s Developer Console, Business Manager, API documentation and a code editor, developers can describe what they want to set up to an AI agent, which can then execute the relevant configuration steps.

The tool can create a WhatsApp Business Account, add and verify a phone number through OTP and register the number for the Cloud API. It can also check whether prerequisites such as accepting Meta’s Terms of Service, adding a payment method and completing Business Verification have been met. Where something is still pending, the tool provides links to complete the requirement.

The AI agent’s job does not stop at account creation. During development, it can create message templates from a description, edit existing templates and provide links to check their approval status. Developers can also use it to send test messages from registered numbers and configure callback URLs and field subscriptions for webhooks.

Meta has built authentication checks into the workflow as well. Users sign in through Facebook Login for Business and grant specific permissions before using the tools. The server checks whether the user is an administrator of the app, identifies the associated business and verifies the Terms of Service status before an action is executed.

Reads operate within the user’s own viewer context, while each invocation is logged. For actions that change an account’s state, Meta requires authentication by a person rather than relying on an app-level credential.

The MCP is currently discoverable through AI coding environments including Claude, Codex and ChatGPT, with Cursor also identified among the compatible agents. Access is being rolled out gradually, so the tool may not yet be available to every developer.

Meta’s WhatsApp-specific MCP sits alongside its broader Meta Social Technologies MCP. While the latter is designed for tasks such as finding Graph API endpoints, searching Meta documentation and troubleshooting errors, the WhatsApp Business Tools MCP focuses specifically on accounts, phone numbers, templates, webhooks and test messaging.

The distinction is important because Meta is moving beyond AI that merely writes code. In this setup, an agent can also interact with the tools needed to configure and test the service, turning a conversational request into an actual development workflow.

For now, however, the new MCP is aimed at development and testing rather than production messaging at scale. With the rollout still gradual, Meta is effectively giving AI agents a new set of WhatsApp buttons to press but not yet handing them the keys to the entire production operation.