PayArk MCP Server
Connects AI agents to Nepal's payment infrastructure through the PayArk API. It enables users to monitor transaction history, manage projects, and generate hosted checkout links using natural language.
README
@payark/mcp-server
<p align="center"> <img src="https://raw.githubusercontent.com/Payark-Inc/payark/main/apps/public-demo/public/og-image.png" alt="PayArk MCP Server" width="100%" /> </p>
<p align="center"> <strong>Connect your AI agent to Nepal's payment infrastructure — in minutes.</strong> </p>
What is this?
The PayArk MCP Server is a Model Context Protocol integration that gives AI agents like Claude, Cursor, and Gemini direct access to your PayArk account. Instead of switching between dashboards or writing boilerplate API calls, you can ask your AI in plain language:
"How many payments did we process through Khalti in the last 24 hours?" "Generate a checkout link for Rs. 500." "Why did payment
pay_abc123fail?"
Available Tools
| Tool | What it does |
|---|---|
list_projects |
Lists all your PayArk projects and their IDs |
list_payments |
Fetches transaction history with optional filters |
get_payment |
Gets the full state of a single payment |
create_payment_link |
Generates a hosted checkout URL |
Quick Start
Step 1 — Get your credentials
- Go to your PayArk Dashboard → Developers.
- Generate a Personal Access Token (PAT).
- Keep it secret — it grants account-level access.
Step 2 — Connect your agent
Claude Desktop
Add this to your claude_desktop_config.json:
{
"mcpServers": {
"payark": {
"command": "npx",
"args": ["-y", "@payark/mcp-server"],
"env": {
"PAYARK_PAT": "pat_your_token_here",
"PAYARK_API_URL": "https://api.payark.com"
}
}
}
}
Cursor (Desktop IDE)
- Open Cursor Settings → MCP.
- Click + Add New MCP Server.
- Set the name to
PayArk, type tocommand, and use this command:
export PAYARK_PAT=pat_... && npx @payark/mcp-server
Remote / SSE Mode
For web-based agents or cloud-to-cloud connections, run the server in SSE mode:
export PAYARK_PAT=pat_...
export PORT=3001
bun run @payark/mcp-server --sse
The SSE endpoint will be available at http://localhost:3001/sse.
Security
Authentication is handled via Personal Access Tokens (PATs), which are scoped to your user account rather than any single project. This means:
- Full audit trail — every MCP-triggered action appears in your account's audit log.
- Easy revocation — if a token is compromised, revoke it instantly from the dashboard.
- Encrypted in transit — all communication with the PayArk API uses TLS 1.3.
Prompting Tips
To get better results from your agent, give it context upfront:
"You are a FinTech assistant with access to PayArk MCP tools. Help me monitor transactions and generate payment links. When I ask for a payment link, suggest the best provider based on recent success rates."
License
MIT © 2026 PayArk Labs.
推荐服务器
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。
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。
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。
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。
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。
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。
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。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
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。