Skip to content
Model Context Protocol (MCP)

Connect WhatsApp to Any AI

Talk to your WhatsApp from Claude, ChatGPT, VS Code Copilot, Cursor, Gemini and more. Send messages, analyze conversations, and automate WhatsApp using natural language.

Get Your API KeyView Setup Guide
claude-desktop
// Asking Claude to send a WhatsApp message through Wassenger

User: Send a message to John saying
      "Meeting confirmed for tomorrow at 3 PM"

Claude: I'll send that WhatsApp message to John right away.

✓ Message sent successfully
Compatible

Works with Your Favorite AI Tools

Connect Wassenger to any MCP-compatible AI assistant or development tool.

  • Claude
  • ChatGPT
  • VS Code Copilot
  • Cursor
  • Windsurf
  • Gemini
  • Zed Editor
  • Continue.dev
  • Aider
Natural language

Talk to WhatsApp in Natural Language

Just tell your AI what you want to do with WhatsApp. No code required.

Messaging

  • Send a message to John saying meeting confirmed at 3 PM.
  • Forward the latest invoice from accounting to all suppliers.
  • Reply to Sara that the order has been shipped.

Analysis

  • Summarise the last 50 messages from the sales channel.
  • Find every chat where a customer mentioned "refund" this week.
  • Show me the busiest contacts in the past 7 days.

Management

  • Assign all unanswered chats to the support team.
  • Archive every chat from before April 1.
  • Add the label "VIP" to my top 10 buyers.

Campaigns

  • Schedule a campaign for tomorrow at 10 AM to all premium customers.
  • Send the new product launch to the "Subscribers" list.
  • Re-engage every contact that hasn't replied in 30 days.

Contacts

  • Find every contact tagged "lead" in Spain.
  • Update Alex's phone to +34 612 345 678.
  • Export all "VIP" contacts as a CSV.

Automation

  • Run the welcome flow on every new contact this morning.
  • Trigger the abandoned-cart flow for the last 50 orders.
  • Pause every active automation until Monday.
Setup

Quick Setup

Add your MCP server URL to your favorite AI tool and start chatting with WhatsApp.

MCP Server URL
https://api.wassenger.com/mcp?key=YOUR_API_KEY

Replace YOUR_API_KEY with your Wassenger API key. Get one from the dashboard.

claude_desktop_config.json
{
  "mcpServers": {
    "wassenger": {
      "type": "http",
      "url": "https://api.wassenger.com/mcp?key=YOUR_API_KEY"
    }
  }
}
Tools

20+ WhatsApp Tools at Your Fingertips

Full access to WhatsApp capabilities through the MCP protocol.

Messages

Send text, media, location, contacts, polls, reactions and more.

send_whatsapp_messageget_whatsapp_chat_messagesanalyze_whatsapp_chat_messages

Chats

List, archive, assign and search chats across your inbox.

get_whatsapp_chatsget_whatsapp_unread_chatssearch_whatsapp_chats_by_name

Groups

Manage WhatsApp groups, post to channels and add participants.

manage_whatsapp_groupsmanage_whatsapp_group_participantsmanage_whatsapp_channels

Campaigns

Create, schedule and monitor bulk WhatsApp campaigns.

manage_whatsapp_campaignsmanage_whatsapp_campaign_contactslist_whatsapp_templates

Contacts

Search contacts, update labels and export segments.

manage_whatsapp_labelsvalidatePhoneNumbersverifyWhatsAppNumberExists

Devices & Team

Inspect device status, queue health and team assignments.

get_whatsapp_devicesmanage_whatsapp_teammanage_whatsapp_queue
Why MCP

Why Use MCP for WhatsApp?

One open standard, every AI assistant. Wassenger's MCP server gives you the safest fast-path between your WhatsApp inbox and the AI tools your team already uses.

  • No Local Installation

    Cloud-based MCP server with HTTP streaming. Just add the URL and start.

  • Always Up to Date

    Automatic updates to the latest server version. Zero maintenance on your side.

  • Secure by Default

    Per-workspace API key. The AI client never sees your raw WhatsApp credentials.

  • Human-in-the-Loop

    Send actions wait for your approval by default. Read-only queries run autonomously.

MCP Connection · live
AI Agent
Wassenger MCP
Read inbox · approved
Draft reply · approved
Send to Sara · sent ✓✓
FAQ

Frequently Asked Questions

Everything you need to know about the WhatsApp MCP server.

What is MCP and why does it matter for WhatsApp?
MCP (Model Context Protocol) is an open standard that lets AI assistants like Claude, ChatGPT, Cursor and Gemini call external tools through a single uniform interface. Wassenger ships an official MCP server that exposes your WhatsApp Business inbox as ready-to-use tools — so any MCP-aware AI can read chats, send messages, search contacts and trigger automations on your live WhatsApp number.
Which AI assistants work with the Wassenger MCP server?
Claude Desktop (native MCP support), ChatGPT (via Connectors panel), Cursor and Windsurf (IDE-first), Google Gemini (Vertex AI), and any custom agent built with LangChain, LlamaIndex or the OpenAI / Anthropic APIs. The MCP standard is open, so once Wassenger is connected it works across every compatible client.
Do I need to write code to use MCP with WhatsApp?
No. The whole connection is no-code: open your AI tool's Settings → Connectors panel, find Wassenger, paste your API key and click Connect. Five minutes, one time. Code is only needed if you're building a custom agent from scratch — for everyday use the UI flow is enough.
Is the MCP server self-hosted or managed by Wassenger?
The Wassenger MCP server is fully managed — it runs on our infrastructure and updates automatically when new WhatsApp features ship. There is nothing to install, host or maintain on your side. The same hosted endpoint serves every workspace, isolated by per-workspace API keys.
How does authentication work? Is my WhatsApp data safe?
Each workspace gets its own API key. The AI assistant authenticates against the MCP server using that key on every call. The key never leaves your AI client and Wassenger only accepts calls from the exact workspace the key belongs to. Revoke any key from your account settings at any time — the AI loses access immediately. Wassenger is hosted in the EU and follows GDPR.
Which WhatsApp actions can an AI perform through MCP?
More than 20 tools across Messages (send, read, search), Chats (list, archive, assign), Groups (manage, post), Campaigns (create, schedule), Contacts (search, update), and Devices & Team (status, queue). The AI can read inbound conversations, draft replies, fire automation Flows and pull analytics — all subject to your approval before sending.
Does the AI act autonomously or wait for my approval?
By default the AI shows you what it is about to do — "I will send this message to Alex" — and waits for your green light. You can enable autonomous actions for safe operations (read-only queries, scheduled flows) but writing to WhatsApp always defaults to human-in-the-loop. Configure per-tool approval rules from your workspace settings.
Is the MCP integration available on every plan?
The MCP server for AI agents is included on the Enterprise plan. Professional and Business plans can still use Wassenger via the REST API and webhooks, but MCP-native AI tooling (Claude, ChatGPT Connectors, Cursor) requires Enterprise. See the pricing page for the full plan comparison.

Ready to Connect WhatsApp to AI?

Get your API key and start automating WhatsApp with natural language in minutes.

Get Started FreeRead Documentation

For developers

Full REST API and OpenAPI spec, signed webhooks, official SDKs in 6 languages plus the MCP server.

API documentationMCP server on GitHubSDKs