Antom MCP Server
Official# Antom MCP Server
A Model Context Protocol (MCP) compatible server that integrates Ant International's Antom payment APIs, enabling AI assistants to handle payment and refund operations seamlessly.
## Overview
The Antom MCP Server wraps Ant International's Antom payment APIs into standardized MCP tools, allowing AI assistants to securely process payment-related operations during conversations. With this server, you can create payment sessions, query transaction status, handle refunds, and more directly through AI interactions.
## Features
### 💳 Payment Operations
- **Create Payment Session** (`create_payment_session`): Generate payment sessions for client-side SDK integration
- **Query Payment Details** (`query_payment_detail`): Retrieve transaction status and information for submitted payment requests
- **Cancel Payment** (`cancel_payment`): Cancel payments when results are not returned within expected timeframes
### 💰 Refund Operations
- **Create Refund** (`create_refund`): Initiate full or partial refunds against successful payments
- **Query Refund Details** (`query_refund_detail`): Check refund status for previously submitted refund requests
## Prerequisites
Before using the Antom MCP Server, ensure you have:
- **Python 3.11 or higher**
- **uv** (recommended package manager) or **pip**
- **Valid Antom Merchant Account** with:
- Merchant Client ID (CLIENT_ID)
- Merchant RSA Private Key (MERCHANT_PRIVATE_KEY)
- Alipay RSA Public Key (ALIPAY_PUBLIC_KEY)
- Payment Redirect Return URL (PAYMENT_REDIRECT_URL)
- Payment Notification Callback URL (PAYMENT_NOTIFY_URL)
## Quick Start
### 1. Installation
#### Direct Usage with uvx (Recommended)
```bash
uvx ant-intl-antom-mcp
```
#### Install from Source
```shell
git clone https://github.com/alipay/global-antom-mcp.git
cd global-antom-mcp
uv install
```
### 2. MCP Client Configuration
Add the following configuration to your MCP client:
```json
{
"mcpServers": {
"antom-mcp-server": {
"command": "uvx",
"args": ["ant-intl-antom-mcp"],
"env": {
"GATEWAY_URL": "https://open-sea-global.alipay.com",
"CLIENT_ID": "your_client_id_here",
"MERCHANT_PRIVATE_KEY": "your_merchant_private_key_here",
"ALIPAY_PUBLIC_KEY": "your_alipay_public_key_here",
"PAYMENT_REDIRECT_URL": "/",
"PAYMENT_NOTIFY_URL": "https://your-domain.com/payment/notify"
}
}
}
}
```
### 3. Environment Variables
| Variable | Required | Description |
| --- |----------|------------------------------------------------------------------------|
| `GATEWAY_URL` | ❌ | Antom API gateway URL (defaults to https://open-sea-global.alipay.com) |
| `CLIENT_ID` | ✅ | Merchant client ID for identity verification |
| `MERCHANT_PRIVATE_KEY` | ✅ | Merchant RSA private key for request signing |
| `ALIPAY_PUBLIC_KEY` | ✅ | Alipay RSA public key for response verification |
| `PAYMENT_REDIRECT_URL` | ❌ | The user is redirected to after the payment is completed |
| `PAYMENT_NOTIFY_URL` | ❌ | Payment result notification callback URL |
## Integration Example
Here's how you can integrate the Antom MCP Server with your AI agent (using QwenAgent as an example):
```python
import os
from qwen_agent.agents import Assistant
# Configure the MCP server as a tool
tools = [{
"mcpServers": {
"antom-mcp-server": {
"command": "uvx",
"args": ["ant-intl-antom-mcp"],
"env": {
"CLIENT_ID": os.getenv('CLIENT_ID'),
"MERCHANT_PRIVATE_KEY": os.getenv('MERCHANT_PRIVATE_KEY'),
"ALIPAY_PUBLIC_KEY": os.getenv('ALIPAY_PUBLIC_KEY'),
"GATEWAY_URL": "https://open-sea-global.alipay.com",
"PAYMENT_REDIRECT_URL": "/",
"PAYMENT_NOTIFY_URL": "https://your-domain.com/notify"
}
}
}
}]
# Create your AI assistant with payment capabilities
bot = Assistant(
llm={'model': 'qwen-max', 'api_key': 'your-api-key'},
function_list=tools,
system_message="You are a helpful assistant with payment processing capabilities."
)
```
## Changelog
See [CHANGELOG.md](CHANGELOG.md) for a detailed history of changes.
## License
This project is licensed under the MIT License.
## Acknowledgments
- [Model Context Protocol](https://modelcontextprotocol.io/) for the standard
- [Antom Integration](https://docs.antom.com/ac/cashierpay/quick_start?platform=Web&client=HTML&server=Python&integration_type=CKP-HOSTED) for the Antom payment platform
TDQS
Scored across 5 tools
Each tool handles a distinct payment operation: creating a payment session, canceling a pending payment, creating a refund, querying payment details, and querying refund details. There is no overlap in functionality.
All tool names follow a consistent verb_noun pattern in snake_case (e.g., cancel_payment, create_payment_session, create_refund). This provides a predictable and clear naming convention.
With 5 tools, the server is well-scoped for payment processing. The count is neither too few nor too many, covering core payment lifecycle operations without unnecessary clutter.
The tool set covers the essential payment operations: initiation, cancellation, refund, and status queries for both payments and refunds. No obvious gaps exist for a typical payment processing workflow.