chatkazi-mcp
Bridges MCP clients with ChatKazi WhatsApp API, enabling AI models to manage WhatsApp sessions, monitor connection states, and send text/media messages.
README
ChatKazi MCP
An implementation of the Model Context Protocol (MCP) designed to act as a bridge between an MCP-compliant client and the ChatKazi WhatsApp API gateway.
This server exposes a series of tools that enable an AI model to interact with WhatsApp instances, programmatically managing sessions, monitoring connection states, and dispatching text and media messages.
Architectural Overview
This server functions as a local translation layer running over Standard Input/Output (Stdio) transport:
[ MCP Client / Host ] <--- Stdio (JSON-RPC) ---> [ chatkazi-mcp Server ] <--- HTTPS (fetch) ---> [ ChatKazi API ]
- Client / Host Handshake: The host launches this server as a subprocess. The server registers its capabilities and schemas.
- Dynamic Tool Execution: When the AI model decides to call a tool, the host routes a JSON-RPC request over Stdio.
- API Forwarding: The server translates the arguments, formats the HTTPS headers, executes the fetch call to ChatKazi, parses the result, and returns a clean text context block to the model.
Environment Configuration
The server relies on the host environment to supply authentication variables. These must be defined when spawning the process:
CHATKAZI_API_KEY(Required): Your live API key retrieved from your ChatKazi dashboard.CHATKAZI_BASE_URL(Optional): The base URL of the API. Defaults tohttps://api.chatkazi.app/api/v1if omitted.
WhatsApp Session Lifecycle
The AI model can track and manage the lifecycle of connected WhatsApp numbers. Under the hood, sessions transition through the following states, which are returned by the session status tools:
| State | Meaning | Required Action |
|---|---|---|
not_started |
The session has been created but not initialized. | Call start_session |
connecting |
ChatKazi is preparing the connection. | Wait and poll status |
qr_ready |
A QR code pairing link is active. | Retrieve QR and scan in WhatsApp |
pairing_code_ready |
A text pairing code has been generated. | Retrieve code |
connected |
The session is online and ready. | Proceed to send messages |
disconnected |
The connection dropped or failed. | Re-authenticate or restart session |
logged_out |
The session was explicitly terminated. | Run start_session to reconnect |
Detailed Tool Reference
The following tools are dynamically exposed to the AI model:
Session Management Tools
list_sessions
- Description: Retrieve a list of all WhatsApp session instances owned by the ChatKazi account.
- Parameters: None.
start_session
- Description: Prepares and starts a session instance.
- Parameters:
sessionId(string, optional): A unique identifier. Defaults to"default".
get_session_status
- Description: Queries the current state of a session.
- Parameters:
sessionId(string, required): The target session identifier.
get_session_qr
- Description: Retrieves the QR code link to link a device.
- Parameters:
sessionId(string, required): The target session identifier.
logout_session
- Description: Logs out and disconnects a session, clearing all state.
- Parameters:
sessionId(string, required): The target session identifier.
Messaging Tools
send_whatsapp_text
- Description: Dispatches a plain text WhatsApp message.
- Parameters:
to(string, required): Recipient number in international format without spaces or+(e.g.,"254712345678").text(string, required): The text content of your message.sessionId(string, optional): The session identifier to send from. Defaults to"default".
send_whatsapp_media
- Description: Dispatches an image, video, audio, or document file.
- Parameters:
to(string, required): Recipient number in international format without spaces or+.url(string, required): Publicly reachable URL of the direct file asset (e.g., a direct link ending in.pngor.jpg).type(string, optional): One ofimage,video,audio, ordocument. Defaults toimage.caption(string, optional): Optional caption text.fileName(string, optional): Display name for document attachments.mimetype(string, optional): Explicit MIME type.sessionId(string, optional): The session identifier to send from. Defaults to"default".
Error Mapping & Resilience
When the ChatKazi API returns an error, the server intercepts the HTTP status and surfaces semantic guidance to the AI model rather than throwing unhandled exceptions. This allows the model to alter its behavior and advise the user:
- HTTP 401 (Unauthorized): Surfaced as a credential configuration error, advising the user to check their
CHATKAZI_API_KEY. - HTTP 402 (Payment Required / Limit Reached): Surfaced as a plan limitation error (e.g., "WhatsApp instance limit reached"), informing the model to stop trying to create instances or sending messages and advise the user to upgrade their plan.
- HTTP 409 (Conflict): Surfaced when trying to start an active or conflicted session.
- Empty Media Payload: Intercepts files that fail to download (e.g., file sizes of
0bytes) and warns the model that the URL is likely invalid or blocked.
General Integration (MCP Client Setup)
To integrate this server into an MCP client or host platform, add the server to your host's configuration settings:
{
"mcpServers": {
"chatkazi-mcp": {
"command": "npx",
"args": [
"-y",
"@lxmwaniky/chatkazi-mcp"
],
"env": {
"CHATKAZI_API_KEY": "your_chatkazi_api_key_here",
"CHATKAZI_BASE_URL": "https://api.chatkazi.app/api/v1"
}
}
}
}
Client Integration Guides
Find instructions below to connect this MCP server to your favorite developer tool.
<details> <summary><b>1. Claude Desktop</b></summary>
Add the configuration to your claude_desktop_config.json:
- File Path (macOS):
~/Library/Application Support/Claude/claude_desktop_config.json - File Path (Windows):
%APPDATA%\Claude\claude_desktop_config.json
{
"mcpServers": {
"chatkazi-mcp": {
"command": "npx",
"args": [
"-y",
"@lxmwaniky/chatkazi-mcp"
],
"env": {
"CHATKAZI_API_KEY": "YOUR_CHATKAZI_API_KEY_HERE"
}
}
}
}
</details>
<details> <summary><b>2. Claude Code (CLI)</b></summary>
Dynamically register this server with the Claude CLI:
claude mcp add chatkazi-mcp npx -y @lxmwaniky/chatkazi-mcp --env CHATKAZI_API_KEY="YOUR_CHATKAZI_API_KEY_HERE"
</details>
<details> <summary><b>3. Cursor</b></summary>
Connect this server visually in your editor settings:
- Open Cursor Settings > Features > MCP.
- Click + Add New MCP Server.
- Configure the following fields:
- Name:
chatkazi-mcp - Type:
command - Command:
npx -y @lxmwaniky/chatkazi-mcp
- Name:
- Click + Add Env to configure your
CHATKAZI_API_KEY. </details>
<details> <summary><b>4. Windsurf</b></summary>
Add this to your Codeium MCP server configuration file:
- File Path:
~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"chatkazi-mcp": {
"command": "npx",
"args": [
"-y",
"@lxmwaniky/chatkazi-mcp"
],
"env": {
"CHATKAZI_API_KEY": "YOUR_CHATKAZI_API_KEY_HERE"
}
}
}
}
</details>
<details> <summary><b>5. Antigravity CLI</b></summary>
Configure this server directly inside your central configuration:
"chatkazi-mcp": {
"command": "npx",
"args": [
"-y",
"@lxmwaniky/chatkazi-mcp"
],
"env": {
"CHATKAZI_API_KEY": "YOUR_CHATKAZI_API_KEY_HERE"
}
}
</details>
Local Development
To compile and verify this project locally:
Prerequisites
- Node.js (v18 or higher is recommended)
- pnpm (preferred package manager)
Step-by-Step Setup
-
Install dependencies: Installs node modules and setup definitions:
pnpm install -
Compile TypeScript code: Compiles the TypeScript source files from the
src/directory into JavaScript inside thebuild/directory:pnpm run build -
Grant executable permissions to the output binary: Makes the built JS file directly executable since it starts with a Node shebang:
chmod +x build/index.js -
Verify using the MCP Inspector: Inspect tools dynamically and test real-time payloads:
CHATKAZI_API_KEY="your_api_key" pnpm dlx @modelcontextprotocol/inspector node build/index.js
Disclaimers
1. WhatsApp Terms of Service (ToS) Compliance
Using third-party, unofficial WhatsApp gateway interfaces carries an inherent risk of account suspension by WhatsApp. WhatsApp strictly prohibits bulk messaging, automated spamming, and unsolicited outreach. You are solely responsible for ensuring your messaging workflows comply with WhatsApp's official Terms of Service. The developers and contributors of this MCP server are not liable for any suspended, blocked, or permanently banned WhatsApp numbers resulting from the use of this software.
2. No Official Affiliation
This project is an independent, community-driven open-source integration. It is not officially affiliated with, endorsed by, sponsored by, or associated with ChatKazi, Meta Platforms, Inc., or WhatsApp. All product names, trademarks, and registered trademarks are the property of their respective owners.
3. Financial and Usage Liability
While this MCP server software is distributed for free under an open-source license, utilizing the ChatKazi API may incur financial costs depending on your ChatKazi subscription tier. You are entirely responsible for monitoring your own platform usage, API call quotas, billing, and any rate limits associated with your ChatKazi account.
4. Warranty and Support
This software is provided "as is," without warranty of any kind, express or implied. The underlying ChatKazi API endpoints, request/response formats, or authentication methods may change over time without notice, which could temporarily or permanently impact the functionality of this server.
License
This project is licensed under the ISC License.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.