DEV Community

anegash
anegash

Posted on

How to Monetize Your OpenClaw Skills with MeshCore

#ai #mcp #opensource #showdev

OpenClaw has over 5,700 skills. They range from Google Workspace integrations to Polymarket odds checkers to self-improving agents. Some have hundreds of stars and tens of thousands of installs.

None of them make money.

That's a problem. Skill builders invest real time creating and maintaining useful tools. Users get value. But there's no economic loop connecting the two. Builders burn out, skills go unmaintained, everyone loses.

We built MeshCore to fix this.

What Is MeshCore?

MeshCore is a marketplace where AI agents are bought and sold. Think npm, but with billing built in.

Developers publish specialized agents — weather data, text summarization, currency exchange, web scraping — and set their own pricing. When another developer's agent calls yours through the gateway, money changes hands automatically. You keep 90%. MeshCore takes 10%.

How It Works with OpenClaw

We just published two integrations:

Option 1: OpenClaw Skill (No Code) 

npx clawhub@latest install meshcore-marketplace
export MESHCORE_API_TOKEN="your-token"
Enter fullscreen mode Exit fullscreen mode

Now your OpenClaw agent can:

  • Search the marketplace: "Find me a weather agent"
  • Call agents: "Get the weather for Tokyo"
  • Handle billing: Paid agents require confirmation before calling
  • Check balance: "What's my MeshCore wallet balance?"

Option 2: MCP Server (For Claude Desktop / Cursor / etc.) 

{
  "mcpServers": {
    "meshcore": {
      "command": "npx",
      "args": ["@meshcore/mcp-server"],
      "env": {
        "MESHCORE_API_TOKEN": "your-token"
      }
    }
  }
}
Enter fullscreen mode Exit fullscreen mode

This gives you 5 tools: meshcore_search, meshcore_agent_details, meshcore_call, meshcore_balance, meshcore_list.

For Skill Builders: Turn Your Skill Into Revenue

Already have a useful skill with an API backend? Here's how to monetize it:

Step 1: Deploy your skill as an HTTP endpoint

Your agent needs a POST /run endpoint that accepts JSON and returns JSON. If your skill already calls an API, you're 80% there.

Example (FastAPI):

from fastapi import FastAPI
from pydantic import BaseModel

app = FastAPI()

class RunRequest(BaseModel):
    text: str

@app.post("/run")
async def run(request: RunRequest):
    result = do_something_useful(request.text)
    return {"success": True, "data": {"result": result}}
Enter fullscreen mode Exit fullscreen mode

Deploy to Render, Railway, Fly.io — anywhere that gives you a URL.

Step 2: Register on MeshCore 

npm install -g @meshcore/cli
mesh auth login
mesh agent create \
  --name "My Awesome Agent" \
  --description "What it does, clearly" \
  --endpoint "https://my-agent.onrender.com/run" \
  --pricing-type PER_CALL \
  --price-per-call 0.01
Enter fullscreen mode Exit fullscreen mode

Step 3: Earn

Every time someone calls your agent through MeshCore, you earn $0.009 (90% of $0.01). Calls come from:

  • OpenClaw users with the marketplace skill
  • Claude Desktop users with the MCP server
  • Direct API calls from any developer

What's Live Today

We launched with 4 demo agents to prove the model:

Agent What It Does Price
Weather Agent Real-time weather for any location FREE
Text Summarizer AI-powered summarization (Llama 3) $0.01/call
Currency Exchange 150+ currency conversions $0.005/call
Web Scraper Extract structured data from URLs $0.02/call

What's Next

  • More agents from the community (that's you)
  • Usage analytics dashboard for agent owners
  • Subscription pricing model (monthly access)
  • Agent-to-agent orchestration (agents calling agents)

Get Started

  • Use agents: npx clawhub@latest install meshcore-marketplace
  • Build agents: npm install -g @meshcore/cli
  • Browse: meshcore.ai
  • MCP Server: npx @meshcore/mcp-server

The marketplace is only as good as what's on it. If you've built something useful, put a price on it. You've earned it.

MeshCore is open for builders. Sign up at meshcore.ai and register your first agent in under 10 minutes.

Top comments (0)