Costa Rica Payments MCP — let AI agents accept payments in Costa Rica

Use it in 60 seconds

Paste this into your MCP client config (Claude Desktop, Cursor, Windsurf, or any MCP-capable agent):

{
  "mcpServers": {
    "costa-rica-payments": {
      "type": "http",
      "url": "https://mcp-cr.wishpool.app/mcp"
    }
  }
}

Nothing to install. Credentials, when you need them, travel as HTTP headers on each request and are never stored — see the threat model.

Or run it yourself

Would you rather not send production credentials to a server you do not control? Deploy this identical code to your own account and point your agent at your own URL:

Deploy with Vercel

git clone https://github.com/junter1989k-ai/costa-rica-payments-mcp && cd costa-rica-payments-mcp && npx vercel --prod

MIT-licensed. Self-hosting removes us from the picture entirely, at no cost and with no loss of function.

---

Live · MCP endpoint: https://mcp-cr.wishpool.app/mcp · Listed on the official MCP Registry as app.wishpool/costa-rica-payments-mcp · Part of a 81-country family — hub: mcp.wishpool.app

A remote MCP server that lets any AI agent (Claude, ChatGPT, Cursor…) accept payments in Costa Rica: cards, SINPE Movil — via ONVO Pay — via ONVO Pay's hosted checkout.

⚖️ Stateless translation layer: funds always flow buyer → ONVO Pay → merchant directly. Never holds funds or stores credentials — no database at all.

Quick start

MCP endpoint: https://mcp-cr.wishpool.app/mcp (Streamable HTTP, stateless JSON-RPC)

Header	Value
x-onvo-secret-key	your ONVO Pay secret key from dashboard.onvopay.com (test keys = test mode)

Tools

Tool	What it does
create_payment_link	Create a CRC hosted-checkout link.
query_payment_status	Pull-based status check; paid=true when PAID.

Develop

node test/serve.js      # local server (http://localhost:3288/mcp)
node test/e2e.js        # protocol + validation suite (probes real ONVO Pay auth errors)

Zero runtime dependencies; Node ≥ 18. MIT license.