Back to Home

Model Context Protocol (MCP)

v1.0

SentinelPay supports the official Model Context Protocol. This allows AI agents like Claude Desktop to automatically execute onboarding tasks for you, without touching a web UI.

Launch Interactive Simulator

How it works

The SentinelPay MCP server exposes a local tool called submit_merchant_application to your AI agent. When you chat with Claude and ask it to onboard a merchant, Claude will automatically parse your prompt, format the data into the correct JSON structure, and securely send it directly to our backend API.

Setup Instructions (Claude Desktop)

Open your Claude Desktop configuration file and add the following JSON block. Note that the Python script must be executed on your local machine.

"mcpServers": {
  "sentinelpay": {
    "command": "python",
    "args": [
      "/Absolute/Path/To/Your/Local/Repo/backend/mcp_server.py"
    ],
    "env": {
      "SENTINELPAY_API_URL": "https://razorpay-onboardingagent.onrender.com"
    }
  }
}

Important details for the config:

  • args: You must replace /Absolute/Path/To/... with the real path where you cloned this repository. Claude needs to know exactly where the Python script lives on your hard drive.
  • env: You do NOT need to modify any local .env files. Claude will automatically inject the `SENTINELPAY_API_URL` directly into the script when it runs it.
  • Dependencies: Ensure your global python environment has the mcp and httpx packages installed (`pip install mcp httpx`).