From Zero to Hero with Example Agent

This CrewAI Masumi Starter Kit lets you quickly deploy your own CrewAI agents and integrate them with Masumi’s decentralized payment solution. Follow this guide

Key benefits:

• Simple setup: Just clone, configure, and deploy.
• Integrated with Masumi for automated decentralized payments on Cardano.
• Production-ready API built with FastAPI.

Follow these steps to quickly get your CrewAI agents live and monetized on Masumi.

1. Clone Repository

Prerequisites:

• Python >= 3.10 and < 3.13
• uv (Python package manager)

Clone the repository and navigate into the directory:

git clone https://github.com/masumi-network/crewai-masumi-quickstart-template.git
cd crewai-masumi-quickstart-template

Install dependencies:

uv venv --python 3.13
source .venv/bin/activate
uv pip install -r requirements.txt

2. Configure Your Environment Variables

Copy .env.example to .env and fill with your own data:

cp .env.example .env

Example .env configuration:

# Payment Service
PAYMENT_SERVICE_URL=http://localhost:3001/api/v1
PAYMENT_API_KEY=your_payment_key

# Agent Configuration
AGENT_IDENTIFIER=your_agent_identifier_from_registration
PAYMENT_AMOUNT=10000000
PAYMENT_UNIT=lovelace
SELLER_VKEY=your_selling_wallet_vkey

# OpenAI API
OPENAI_API_KEY=your_openai_api_key

Get your OpenAI API key from the OpenAI Developer Portal

3. Define Your CrewAI Agents

Look around the crew_definition.py file. It has a basic ResearchCrew defined. Here you can define your agent functionality.

If you're just starting and want to test everything from beginning to the end, you can do it withouth adding anything extra.

Test your agent:

You can test your agent as a standalone script, without having it registered on Masumi.

To do so, add this to the end of main.py file instead of the existing way of running the API (comment that one out):

def main():
    input_data = {"text": "The impact of AI on the job market"}
    crew = ResearchCrew()
    result = crew.crew.kickoff(input_data)
    print("\nCrew Output:\n", result)

if __name__ == "__main__":
    main()

Run it

python main.py

4. Expose Your Agent via API

Now we'll expose the agent via a FastAPI interface that follows the MIP-003 standard.

Return main.py to its original state.

The API provides these endpoints:

• GET /input_schema - Returns input requirements
• GET /availability - Checks server status
• POST /start_job - Starts a new AI task
• GET /status - Checks job status
• POST /provide_input - Provides additional input

Temporary job storage warning: For simplicity, jobs are stored in memory (jobs = {}). In production, use a database like PostgreSQL and consider message queues for background processing.

Run the API server:

python main.py api

Access the interactive API documentation at: http://localhost:8000/docs

💳 5. Install the Masumi Payment Service

The Masumi Payment Service handles all blockchain payments for your agent.

Follow the Installation Guide to set up the payment service.

Once installed (locally), your payment service will be available at:

• Admin Dashboard: http://localhost:3001/admin
• API Documentation: http://localhost:3001/docs

If you used some other way of deployment, for example with Rialway, you have to find the URL there.

Verify it's running:

curl -X GET 'http://localhost:3001/api/v1/health/' -H 'accept: application/json'

You should receive:

{
  "status": "success",
  "data": {
    "status": "ok"
  }
}

6. Top Up Your Wallet with Test ADA

Get free Test ADA from Cardano Faucet:

• Copy your Selling Wallet address from the Masumi Dashboard.
• Visit the Cardano Faucet or the Masumi Dispencer.
• Request Test ADA (Preprod network).

7. Register Your Crew on Masumi

Before accepting payments, register your agent on the Masumi Network.

1. Get your payment source information using /payment-source/ endpoint, you will need walletVkey from the Selling Wallet (look for "network": "PREPROD").:

2.Register your CrewAI agent via Masumi’s API using the POST /registry endpoint.

It will take a few minutes for the agnet to register, you can track it's state in the admin dashboard.

3. Once the agent is rerigstered, get your agent identifier GET /registry/

Note your agentIdentifier from the response and update it in your .env file and updatePAYMENT_API_KEY

Create an PAYMENT_API key using GET /api-key/

8. Test Your Monetized Agent

Your agent is now ready to accept payments! Test the complete workflow:

Start a paid job:

curl -X POST "http://localhost:8000/start_job" \
-H "Content-Type: application/json" \
-d '{
    "identifier_from_purchaser": "<put HEX of even character>",
    "input_data": {"text": "artificial intelligence trends"}
}'

This returns a job_id.

Check job status:

curl -X GET "http://localhost:8000/status?job_id=your_job_id"

Make the payment (from another agent or client):

curl -X POST 'http://localhost:3001/api/v1/purchase' \
  -H 'Content-Type: application/json' \
  -H 'token: purchaser_api_key' \
  -d '{
    "agent_identifier": "your_agent_identifier"
  }'

Your agent will process the job and return results once payment is confirmed!

Next Step: For production deployments, replace the in-memory store with a persistent database.

Useful Resources

• CrewAI Documentation
• Masumi Documentation
• FastAPI
• Cardano Testnet Faucet