Provider Guide·~8 min read

Deploy a paid MCP server
in 60 minutes

Build, deploy, and monetise an MCP server on RaegentAI — from first line of code to live paid calls.

What you'll need

  • ✓Node.js 18+ and npm
  • ✓A publicly reachable HTTPS endpoint (or the ability to deploy one)
  • ✓A RaegentAI account — free to create

What is an MCP server?

The Model Context Protocol (MCP) is an open standard for connecting AI models to external tools and data sources. An MCP server exposes one or more tools — discrete, typed functions that an AI client can discover and call by name. When you list an MCP server on RaegentAI, developers can rent access to those tools through the platform without ever touching your infrastructure directly. You stay in control; we handle routing, billing, and access management.

Step 01

Scaffold your server

Install the official MCP SDK and Zod for input validation. The example below implements a live currency exchange-rate lookup — a realistic tool that calls a public API and returns a text result.

bash
npm install @modelcontextprotocol/sdk zod
server.ts — TypeScript
// server.ts
import { McpServer } from '@modelcontextprotocol/sdk/server/mcp.js'
import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js'
import { z } from 'zod'

const server = new McpServer({ name: 'exchange-rates', version: '1.0.0' })

server.tool(
  'get_exchange_rate',
  { from: z.string(), to: z.string() },
  async ({ from, to }) => {
    const res = await fetch(
      `https://api.frankfurter.app/latest?from=${from}&to=${to}`
    )
    const data = await res.json()
    return {
      content: [{ type: 'text', text: `1 ${from} = ${data.rates[to]} ${to}` }],
    }
  }
)

const transport = new StdioServerTransport()
await server.connect(transport)

The server.tool() call registers a typed tool. The first argument is the tool name agents will call, the second is the Zod input schema (validated automatically before your handler runs), and the third is the async handler itself. Add as many tools as you need before calling server.connect().

Step 02

Deploy it

Your server must be reachable at a public HTTPS URL. RaegentAI pings your endpoint during validation and routes live traffic through it once approved. Any platform works — pick the one you already know.

VercelRecommended

One-command deploy, zero infra config. Works with Express, Fastify, or a Next.js API route.

Railway

Docker-friendly. Set a start command and Railway assigns a public HTTPS URL automatically.

Any HTTPS endpoint

Already running a server? Point RaegentAI at your existing URL — no migration needed.

bash — Vercel
# install Vercel CLI if you haven't already
npm install -g vercel

# deploy from your project root
vercel --prod

After deployment, copy the HTTPS URL — you'll paste it into the submission form in the next step. The URL must remain live and responsive after submission; downtime during validation will cause a rejection.

Step 03

Submit to RaegentAI

Go to raegentai.com/submit-mcp and fill in the form. Here's what each field does:

Server URLThe public HTTPS URL from Step 2. Must respond to POST requests with valid MCP JSON.
NameWhat renters see in search results. Short, descriptive, and unique — not generic like "My Server".
DescriptionA clear explanation of what your tools do and when to use them. This is your main SEO surface.
CategoryPick the closest match from the dropdown. This controls which category pages your server appears on.
Credits per callHow many credits are charged per tool invocation. 1,000 credits = £1. You set this value in Step 4.

Screenshot: submit form

Open the submit form →
Step 04

Set your pricing

Pricing is denominated in credits. 1,000 credits = £1. You set a default_credits_per_call value when submitting — renters pay that amount each time one of your tools is invoked. Pick a number that covers your underlying API costs plus a margin.

Common pricing examples
50£0.05Lightweight lookup — exchange rates, simple data fetch
200£0.20Standard API call with moderate compute or third-party cost
500£0.50LLM-backed enrichment or multi-step processing
2,000£2.00Heavyweight batch operation or expensive third-party API

Revenue split

Free tier
You keep 80%Platform 20%
Pro — £20/mo
You keep 85%Platform 15%

Upgrading to Pro costs £20/mo and raises your take from 80% to 85%. The extra 5% pays back within a few hundred pounds of monthly revenue. See the full breakdown on the pricing page.

Step 05

Test it

Once your submission is approved, you'll receive a test token scoped to your server. Use the Playground at /playground for a browser-based interface, or call the gateway directly:

bash — test your tool via the gateway
curl -X POST https://www.raegentai.com/api/proxy \
  -H "x-raegent-token: YOUR_TEST_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "method": "tools/call",
    "params": {
      "name": "get_exchange_rate",
      "arguments": { "from": "USD", "to": "GBP" }
    }
  }'

A successful response returns the content array from your tool handler. A 502 means RaegentAI could not reach your endpoint — check that your server URL is publicly reachable and responding to POST requests. A 401 means the test token is missing or incorrect.

What happens next

Manual review

Every submission goes through a manual review. We test that the endpoint responds correctly, verify the tool descriptions are accurate, and check the server meets our content standards. Expect a decision within 2 business days.

Discovery

Once live, your server appears on the Explore page and in relevant category pages. Ranking is based on call volume, success rate, response time, and renter ratings. A clear, keyword-rich description also helps organic search discoverability.

Earnings

Credits accumulate in your provider account with every successful tool invocation. You can withdraw earnings at any time once your balance reaches £10. Withdrawals are processed to your connected payment method within 5 business days.

Ready to ship?

Submit your MCP server

Takes under 5 minutes to submit. Start earning on the day you're approved.

Submit your MCP server →View pricing →