whatsapp-mcp-server

whatsapp-mcp-server

Connects AI assistants to the official Meta WhatsApp Cloud API for managing conversations and sending various message types through natural language. It provides tools for media management, template messages, and real-time webhook processing without the risk of account bans.

Category
访问服务器

README

whatsapp-mcp-server

License: MIT Node.js TypeScript MCP Compatible Official Meta API WhatsApp Business No Ban Risk

A production-ready Model Context Protocol (MCP) server that connects AI assistants like Claude to the WhatsApp Cloud API (Meta). Send messages, manage conversations, handle media, send template messages, and receive real-time webhooks — all through natural language.


Why this MCP server?

This is the only WhatsApp MCP server built on the official Meta WhatsApp Cloud API.

Most WhatsApp tools for developers rely on unofficial browser automation libraries (whatsapp-web.js) or reverse-engineered protocols (whatsmeow, Baileys). These approaches violate WhatsApp's Terms of Service and carry a real risk of account bans. This project is different.

whatsapp-mcp-server (this project) Unofficial libraries (whatsapp-web.js, whatsmeow, Baileys)
API type ✅ Official Meta WhatsApp Cloud API ❌ Unofficial / reverse-engineered
Account ban risk ✅ None — Meta-approved ⚠️ High — violates WhatsApp ToS
Account type ✅ WhatsApp Business account Personal account only
Uptime reliability ✅ Meta SLA-backed infrastructure ❌ Breaks when WhatsApp updates
Scalability ✅ Business-grade, multi-agent ready ❌ Limited to one session
Template messages ✅ Full support ❌ Not available
Media handling ✅ Native Cloud API media endpoints ⚠️ Workarounds required
Webhook support ✅ Official Meta webhook events ❌ Polling / fragile listeners
Production use ✅ Built for production ❌ Recommended for personal/hobby use only
Long-term viability ✅ Stable, versioned API ❌ Depends on reverse engineering staying current

Bottom line: if you are building anything beyond a personal experiment — customer support, automated notifications, business workflows — use the official API. This server gives you that, packaged as a drop-in MCP server.


Features

  • Send all message types — text, images, videos, audio, documents, reactions
  • Template messages — send and list approved WhatsApp Business templates with full component support
  • Media management — upload, retrieve, and delete media files
  • Conversation history — in-memory store of recent conversations via webhook
  • Real-time webhooks — receive incoming messages and delivery status updates
  • Business profile — read and update your WhatsApp Business profile
  • Contact lookup — validate phone numbers against WhatsApp
  • Rate limit handling — automatic exponential backoff retry logic
  • Type-safe — fully typed TypeScript with Zod input validation
  • Docker ready — multi-stage Dockerfile and docker-compose included
  • Works without webhook — send-only mode with no webhook server required

Architecture

┌─────────────────────────────────────────────────────────────┐
│                    AI Assistant (Claude)                      │
└──────────────────────────┬──────────────────────────────────┘
                           │ MCP Protocol (stdio)
                           │
┌──────────────────────────▼──────────────────────────────────┐
│                  whatsapp-mcp-server                         │
│                                                              │
│  ┌─────────────┐  ┌──────────────┐  ┌───────────────────┐  │
│  │  MCP Server │  │   Tools      │  │  WhatsApp Client  │  │
│  │  (stdio)    │◄─┤  messages    ├─►│  (Axios + retry)  │  │
│  │             │  │  contacts    │  └────────┬──────────┘  │
│  └─────────────┘  │  media       │           │              │
│                   │  templates   │           │ HTTPS        │
│  ┌─────────────┐  └──────────────┘           │              │
│  │  Webhook    │                             ▼              │
│  │  Server     │  ◄──── Incoming   ┌─────────────────────┐ │
│  │  (Express)  │       messages    │ WhatsApp Cloud API  │ │
│  └──────┬──────┘       & statuses  │  graph.facebook.com  │ │
└─────────│───────────────────────────┼─────────────────────┘ │
          │                           │                        │
          │ HTTP POST (Meta webhooks) │ HTTPS (send messages)  │
          │                           │                        │
     ┌────▼──────────────────────────▼──────────────┐
     │            WhatsApp Users                     │
     └────────────────────────────────────────────────┘

Webhook flow for incoming messages:
WhatsApp User → Meta Servers → POST /webhook → WhatsAppWebhook
→ in-memory conversation store → EventEmitter → MCP tools

Prerequisites

  1. Node.js 20+download
  2. Meta Developer Accountcreate one
  3. WhatsApp Business App configured in Meta App Dashboard
  4. Phone Number ID and Access Token from your WhatsApp app settings
  5. (Optional) A public HTTPS URL for receiving webhooks (use ngrok for local dev)

Quick Start

Step 1 — Clone and install:

git clone https://github.com/FredShred7/whatsapp-mcp-server.git
cd whatsapp-mcp-server
npm install

Step 2 — Configure credentials:

cp .env.example .env
# Edit .env with your WhatsApp credentials

Step 3 — Build:

npm run build

Step 4 — Add to Claude Desktop or Claude Code (see sections below)

Step 5 — Test it: Ask Claude: "Send a WhatsApp message to +15551234567 saying Hello from Claude!"


Configuration

All configuration is via environment variables. Copy .env.example to .env and fill in the values.

Variable Required Description
WHATSAPP_PHONE_NUMBER_ID Yes Your WhatsApp phone number ID (from Meta App Dashboard > WhatsApp > API Setup)
WHATSAPP_ACCESS_TOKEN Yes Permanent access token for your app (generate in Meta App Dashboard)
WHATSAPP_WEBHOOK_VERIFY_TOKEN Webhook Random secret string — must match what you enter in Meta App Dashboard
WHATSAPP_BUSINESS_ACCOUNT_ID Optional WhatsApp Business Account ID — required for listing phone numbers and templates
WHATSAPP_API_VERSION Optional WhatsApp Cloud API version (default: v21.0)
WEBHOOK_PORT Optional Port for incoming webhook server (default: off). Set to 3000 to enable
WEBHOOK_PATH Optional HTTP path for webhooks (default: /webhook)
LOG_LEVEL Optional Logging verbosity: info or debug (default: info)

Available Tools

Tool Description Key Parameters
whatsapp_send_text Send a text message to, text, preview_url?, reply_to_message_id?
whatsapp_send_image Send an image with optional caption to, image_url, caption?
whatsapp_send_video Send a video with optional caption to, video_url, caption?
whatsapp_send_audio Send an audio message to, audio_url
whatsapp_send_document Send a document/file to, document_url, caption?, filename?
whatsapp_send_reaction React to a message with emoji to, message_id, emoji
whatsapp_mark_read Mark a message as read message_id
whatsapp_list_conversations List recent conversations (webhook required) limit?
whatsapp_get_conversation Get messages with a contact (webhook required) phone_number, limit?
whatsapp_get_message_status Get delivery/read status (webhook required) message_id
whatsapp_get_contact Validate a phone number on WhatsApp phone_number
whatsapp_get_business_profile Get your WhatsApp Business profile
whatsapp_update_business_profile Update business profile info about?, address?, description?, email?, websites?, vertical?
whatsapp_upload_media Upload media from URL, get media ID media_url, mime_type, filename?
whatsapp_get_media_url Get temporary download URL for media media_id
whatsapp_delete_media Delete uploaded media media_id
whatsapp_send_template Send an approved template message to, template_name, language_code, components?
whatsapp_list_templates List available message templates limit?

Usage with Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):

{
  "mcpServers": {
    "whatsapp": {
      "command": "node",
      "args": ["/absolute/path/to/whatsapp-mcp-server/dist/index.js"],
      "env": {
        "WHATSAPP_PHONE_NUMBER_ID": "your_phone_number_id",
        "WHATSAPP_ACCESS_TOKEN": "your_access_token",
        "WHATSAPP_WEBHOOK_VERIFY_TOKEN": "your_verify_token",
        "WHATSAPP_BUSINESS_ACCOUNT_ID": "your_business_account_id",
        "WEBHOOK_PORT": "3000"
      }
    }
  }
}

Restart Claude Desktop after saving. You should see the WhatsApp tools available in the tools panel.


Usage with Claude Code

Add to your Claude Code MCP settings (~/.claude/settings.json or project-level .claude/settings.json):

{
  "mcpServers": {
    "whatsapp": {
      "command": "node",
      "args": ["/absolute/path/to/whatsapp-mcp-server/dist/index.js"],
      "env": {
        "WHATSAPP_PHONE_NUMBER_ID": "your_phone_number_id",
        "WHATSAPP_ACCESS_TOKEN": "your_access_token",
        "WHATSAPP_WEBHOOK_VERIFY_TOKEN": "your_verify_token",
        "WHATSAPP_BUSINESS_ACCOUNT_ID": "your_business_account_id",
        "WEBHOOK_PORT": "3000"
      }
    }
  }
}

Or use the CLI to add it:

claude mcp add whatsapp node /absolute/path/to/whatsapp-mcp-server/dist/index.js

Webhook Setup

Webhooks allow you to receive incoming messages and delivery status updates. They are optional — the MCP server works without them for sending messages only.

Local development with ngrok

# 1. Install ngrok: https://ngrok.com/download
# 2. Start your webhook server
WEBHOOK_PORT=3000 npm run dev

# 3. In another terminal, expose it publicly
ngrok http 3000

# 4. Copy the HTTPS URL (e.g. https://abc123.ngrok.io)

Configure in Meta App Dashboard

  1. Go to Meta App Dashboard > Your App > WhatsApp > Configuration
  2. Set Callback URL to: https://your-ngrok-url.ngrok.io/webhook
  3. Set Verify Token to match your WHATSAPP_WEBHOOK_VERIFY_TOKEN
  4. Click Verify and Save
  5. Subscribe to the messages webhook field

Production

For production, deploy this server behind a reverse proxy (nginx, Caddy, etc.) with a valid TLS certificate. Set WEBHOOK_PORT to your internal port and expose it via HTTPS on port 443.


Docker Deployment

Build and run with docker-compose

# Copy and fill in your credentials
cp .env.example .env

# Build and start
docker-compose up -d

# View logs
docker-compose logs -f

# Stop
docker-compose down

Build manually

# Build image
docker build -t whatsapp-mcp-server .

# Run (send-only, no webhook)
docker run --rm -i \
  -e WHATSAPP_PHONE_NUMBER_ID=xxx \
  -e WHATSAPP_ACCESS_TOKEN=xxx \
  whatsapp-mcp-server

# Run with webhook server
docker run --rm -i \
  -p 3000:3000 \
  --env-file .env \
  whatsapp-mcp-server

Note: When using Docker with Claude Desktop/Code, the MCP transport uses stdio. Make sure to pass -i (interactive) so stdin/stdout remain connected. The webhook port is separate from the MCP transport.


Development

# Install dependencies
npm install

# Run with auto-reload
npm run dev

# Type-check only (no build)
npm run typecheck

# Build
npm run build

# Lint
npm run lint

Project structure

src/
├── index.ts              # Entry point — loads env, wires up components
├── server.ts             # MCP server — registers tools, handles requests
├── whatsapp/
│   ├── client.ts         # WhatsApp Cloud API client with retry logic
│   ├── types.ts          # TypeScript types and interfaces
│   └── webhook.ts        # Express webhook server + in-memory store
└── tools/
    ├── messages.ts       # Send/receive message tools
    ├── contacts.ts       # Contact and business profile tools
    ├── media.ts          # Media upload/download/delete tools
    └── templates.ts      # Template message tools

Contributing

Contributions are welcome! Please read CONTRIBUTING.md for guidelines on reporting bugs, suggesting features, and submitting pull requests.


License

MIT — Copyright (c) 2025 FredShred7

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选