WhatsApp MCP Server

WhatsApp messages, contacts, and media access via the Model Context Protocol.

Data last scanned today · Reviewed yesterday

4112 tools★ 5,916GoMIT

Overview

This server integrates personal WhatsApp accounts into AI assistants using the whatsmeow library. It enables users to search conversation history, retrieve messages including documents and media, and send new texts to individual or group chats directly from their AI interface. Data is stored in a local SQLite database, with specific content shared only when requested by the AI agent. Users can optionally utilize FFmpeg for automatic audio conversion to ensure compatibility when sending voice messages.

Our verdict

With a high-risk security score of 26/100, this tool requires extreme caution. The server lacks any native authentication, meaning any process with access to your configuration can interact with your account, and the absence of a read-only mode grants the server full write access to your communications. Given these risks and the fact that it has not been updated in over a year, it is recommended only for power users who understand the implications of prompt injection and personal data exfiltration.

Zero direct dependencies reduce the supply-chain attack surface.
Native support for images, videos, documents, and audio.
MIT license provides permissive usage terms.
Local SQLite storage ensures data residency on your machine.

No authentication method implemented for local configuration access.
Lacks a read-only mode to prevent unintended message modifications.
Repository has not received a commit in 370 days.
Community-maintained project without official WhatsApp support.

Tools

Tool	Description	Risk
search_contacts	Search for contacts by name or phone number	low
list_messages	Retrieve messages with optional filters and context	low
list_chats	List available chats with metadata	low
get_chat	Get information about a specific chat	low
get_direct_chat_by_contact	Find a direct chat with a specific contact	low
get_contact_chats	List all chats involving a specific contact	low
get_last_interaction	Get the most recent message with a contact	low
get_message_context	Retrieve context around a specific message	low
send_message	Send a WhatsApp message to a specified phone number or group JID	medium
send_file	Send a file (image, video, raw audio, document) to a specified recipient	medium
send_audio_message	Send an audio file as a WhatsApp voice message (requires the file to be an .ogg opus file or ffmpeg must be installed)	medium
download_media	Download media from a WhatsApp message and get the local file path	low

Compatibility

Client	Local	Docker	Remote	Read-only
Claude Desktop
Cursor
VS Code
Windsurf
ChatGPT

Frequently asked questions

›Do I need any specific software installed to run this server?

Yes, you need Go, Python 3.6+, the UV package manager, and either the Anthropic Claude Desktop app or Cursor. Additionally, FFmpeg is optional but required if you want to automatically convert audio files into playable WhatsApp voice messages.

›How do I authenticate the WhatsApp account?

The authentication is handled by the Go WhatsApp bridge. Upon the first run of the bridge, you will be prompted to scan a QR code using your WhatsApp mobile app.

›Is re-authentication required periodically?

Yes, you may need to re-authenticate the WhatsApp bridge approximately every 20 days.

›Are there specific requirements for running this on Windows?

Yes, you must have a C compiler installed (such as MSYS2) and explicitly enable CGO by running 'go env -w CGO_ENABLED=1' before starting the Go bridge, as go-sqlite3 requires CGO to function.

›Where is my WhatsApp message data stored?

All your message history is stored locally in a SQLite database located within the 'whatsapp-bridge/store/' directory of the repository.

›What are the security implications of using this server?

As with many MCP servers, this project is subject to potential prompt injection risks, which could lead to private data exfiltration if the model is compromised.