How to Earn Money Building MCP Servers (Author Payouts Guide)
You've built an MCP server. It works. Maybe it wraps a useful API, or does something clever with data processing, or gives AI agents a capability they didn't have before.
Now what? It sits in your GitHub repo gathering stars and dust.
What if every time someone used your MCP server, you got paid?
Universal API has a built-in marketplace with author payouts. You publish your MCP server, set your price, and earn revenue every time someone invokes your tools. Payouts happen automatically via Stripe Connect.
This guide covers everything: how to publish, how to price, how to get paid, and what realistic earnings look like.
Last updated: July 2026
How Author Payouts Work
User invokes your MCP server tool
│
▼
Universal API charges the user credits
│
├── Platform fee (infrastructure, compute, billing)
└── Author payout (your share)
│
▼
Stripe Connect Transfer (monthly)
│
▼
Your bank account 💰The flow is simple:
- User pays credits for each invocation of your MCP server
- Platform takes a fee for infrastructure (Lambda compute, API Gateway, S3, billing)
- You keep the rest — your author payout
- Monthly transfer via Stripe Connect to your bank account
Step 1: Publish Your MCP Server
Create your MCP server on Universal API (or migrate an existing one):
const { McpServer } = require("@modelcontextprotocol/sdk/server/mcp.js");
const { z } = require("zod");
function createMcpServer() {
const server = new McpServer({ name: "stock-data", version: "1.0.0" });
server.registerTool("get_stock_price", {
title: "Get Stock Price",
description: "Returns the current stock price for a ticker symbol",
inputSchema: { ticker: z.string().describe("Stock ticker (e.g., AAPL, TSLA)") }
}, async ({ ticker }) => {
// Your implementation here
const data = await fetchStockData(ticker);
return {
content: [{ type: "text", text: JSON.stringify(data) }]
};
});
return server;
}
module.exports = { createMcpServer };Set visibility to public so it appears in the marketplace:
curl -X POST https://api.universalapi.co/mcp-admin/create \
-H "Authorization: Bearer $UNIVERSALAPI_BEARER_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"serverName": "stock-data",
"description": "Real-time stock prices and market data for AI agents",
"sourceCode": "...",
"visibility": "public"
}'Step 2: Set Your Pricing
You control how your server is priced. Universal API supports multiple pricing dimensions:
Per-Invocation Pricing
The simplest model — charge a flat fee per tool call:
{
"authorPricing": {
"pricePerInvocation": 0.001
}
}At $0.001/invocation, if your tool gets called 10,000 times/month, you earn $10/month.
Per-Token Pricing (for LLM-heavy tools)
If your MCP server uses Bedrock models internally, charge based on token usage:
{
"authorPricing": {
"pricePerInputToken": 0.000001,
"pricePerOutputToken": 0.000005
}
}Combined Pricing
Use both for a base fee plus usage:
{
"authorPricing": {
"pricePerInvocation": 0.0005,
"pricePerInputToken": 0.0000005,
"pricePerOutputToken": 0.0000025
}
}Monthly Subscriptions (New)
Instead of (or alongside) per-invocation fees, offer a flat monthly plan. Active subscribers skip your per-invocation fees entirely — platform infrastructure credits still apply:
{
"authorPricing": {
"subscriptionPriceUsdMonthly": 9.99,
"freeTrialQuota": 25
}
}- Minimum price is $0.50/month
- Payment is a Stripe destination charge straight to your Connect account, minus the 20% platform fee
freeTrialQuotagives each consumer N free invocations before they're prompted to subscribe — the single biggest conversion lever
Subscriptions are ideal for heavy-usage tools where per-invocation pricing adds up fast. Consumers subscribe with one click from your server's detail page.
Pricing Strategy Tips
| Strategy | Best For | Example |
|---|---|---|
| Free (no pricing) | Growing user base, open source | Community tools, demos |
| $0.001/invocation | Lightweight tools (search, lookup) | Stock prices, weather, time |
| $0.005/invocation | Medium tools (API wrappers) | Email sending, PDF generation |
| $0.01+/invocation | Heavy tools (compute, LLM) | Document analysis, image gen |
| Per-token | LLM-powered tools | Summarization, translation |
| $5-20/month subscription | Daily-driver tools with heavy usage | Search APIs, data feeds, automation |
Rule of thumb: Price at 2-5x your actual cost. If your tool costs $0.002 in API fees per call, price at $0.005-$0.01.
Step 3: Connect Stripe for Payouts
Set up Stripe Connect to receive payouts:
Via the Dashboard
- Go to Author Dashboard at universalapi.co
- Click "Connect with Stripe"
- Complete Stripe's onboarding (takes ~5 minutes)
- Once connected, payouts happen automatically
Via the API
# Start Stripe Connect onboarding
curl -X POST https://api.universalapi.co/author/stripe/connect \
-H "Authorization: Bearer $UNIVERSALAPI_BEARER_TOKEN"
# Returns: { "data": { "onboardingUrl": "https://connect.stripe.com/..." } }Open the onboardingUrl in your browser to complete onboarding.
Check Your Connect Status
curl https://api.universalapi.co/author/stripe/status \
-H "Authorization: Bearer $UNIVERSALAPI_BEARER_TOKEN"{
"data": {
"connected": true,
"chargesEnabled": true,
"payoutsEnabled": true,
"accountId": "acct_xxx"
}
}Step 4: Track Your Earnings
Author Dashboard
The Author Dashboard at universalapi.co shows:
- My Resources — All your published MCP servers and agents with invocation counts
- Earnings — Revenue breakdown by resource, period, and type
- Payouts — Transfer history with dates, amounts, and status
Earnings API
# Get all-time earnings
curl https://api.universalapi.co/author/earnings \
-H "Authorization: Bearer $UNIVERSALAPI_BEARER_TOKEN"
# Get earnings for a specific month
curl 'https://api.universalapi.co/author/earnings?period=2026-07' \
-H "Authorization: Bearer $UNIVERSALAPI_BEARER_TOKEN"Response:
{
"data": {
"earnings": [
{
"resourceId": "mcp-xxx",
"resourceName": "stock-data",
"invocationCount": 2450,
"totalRevenueUsd": 12.25,
"platformFeeUsd": 4.90,
"authorPayoutUsd": 7.35,
"payoutStatus": "pending"
}
],
"totalAuthorPayoutUsd": 7.35
}
}Payout Schedule
- Frequency: Monthly (1st of each month)
- Minimum payout: $10 USD (earnings below $10 roll over to next month)
- Method: Stripe Transfer to your connected account
- Currency: USD
Example timeline:
- June invocations: $15.00 in earnings
- July 1: Automatic payout of $15.00 via Stripe
- July 2-3: Funds arrive in your bank account
What Realistic Earnings Look Like
Let's be honest about the numbers. MCP tools are cheap per-invocation, so you need volume to make meaningful money.
Scenario 1: Niche Utility Tool
A specialized tool (e.g., PDF invoice parser) with a small but dedicated user base.
- Price: $0.005/invocation
- Users: 20 active users
- Usage: ~50 calls/user/month
- Monthly invocations: 1,000
- Monthly revenue: $5.00
- Your payout: ~$3.00
Verdict: Coffee money. But if your tool is free and you're building reputation, that's fine.
Scenario 2: Popular API Wrapper
A well-built wrapper around a popular API (e.g., Google Search, weather data).
- Price: $0.002/invocation
- Users: 200 active users
- Usage: ~100 calls/user/month
- Monthly invocations: 20,000
- Monthly revenue: $40.00
- Your payout: ~$24.00
Verdict: Covers your API costs with profit.
Scenario 3: High-Value Enterprise Tool
A tool that does something genuinely hard (e.g., document OCR, code analysis).
- Price: $0.02/invocation
- Users: 50 active users
- Usage: ~200 calls/user/month
- Monthly invocations: 10,000
- Monthly revenue: $200.00
- Your payout: ~$120.00
Verdict: Real side income.
Scenario 4: The Dream
A suite of tools that becomes essential for a category of AI agents.
- Price: $0.01/invocation average
- Users: 1,000+ active users
- Monthly invocations: 500,000
- Monthly revenue: $5,000.00
- Your payout: ~$3,000.00
Verdict: This is the goal. It requires building genuinely useful tools and getting distribution.
Tips for Maximizing Earnings
1. Solve a Real Problem
The highest-earning MCP servers aren't clever — they're useful. Tools that save developers time or give agents capabilities they can't get elsewhere:
- API wrappers that handle auth, rate limiting, and error handling
- Data tools that parse, transform, or enrich data
- Integration tools that connect to specific platforms (Jira, GitHub, Salesforce)
2. Write Great Descriptions
Your server's description is how users (and AI agents) find it. Be specific:
- ❌ "A web search tool"
- ✅ "Search Google, Google News, and Reddit via SerpAPI. Returns structured results with titles, URLs, snippets, and timestamps. Supports date filtering and site-specific search."
3. Include Author Keys
If your MCP server wraps a paid API, use author keys (Role Tokens) to include your API credentials. Users don't need their own API keys — they just pay credits. This dramatically increases adoption.
4. Keep It Reliable
Servers that error frequently get abandoned. Test thoroughly, handle edge cases, and return useful error messages.
5. Update Regularly
The AI classification pipeline automatically re-indexes your server when you update it, improving discoverability.
FAQ
Do I need a Stripe account?
Yes — Stripe Connect is how we send payouts. Setting up takes about 5 minutes.
What's the platform fee?
The platform takes a percentage for infrastructure costs (compute, storage, API Gateway, billing). The exact split depends on the resource and its compute costs. Typical author share is 60-70%.
Can I publish for free and switch to paid later?
Yes. Start free to build a user base, then add pricing. Existing users will see the new pricing on their next invocation.
What about agents? Can I earn from published agents too?
Yes — the same author pricing and payout system works for both MCP servers and agents.
Do users see my source code?
By default, public resources show their source code. If you want to keep your implementation private while still allowing public use, contact us about enterprise visibility options.
Getting Started
- Build an MCP server that solves a real problem
- Publish it with
visibility: "public" - Set pricing via the Author Pricing editor
- Connect Stripe in the Author Dashboard
- Promote it — share on Reddit, X, and in relevant communities
- Watch earnings grow in your Author Dashboard
The MCP ecosystem is still early. Building useful tools now — while the marketplace is small — means you'll have the reputation, the reviews, and the distribution when it scales.
Questions about author payouts? Check the Author Dashboard or ask in r/mcp.