Agentic Service Wrapper - CrewAI Branch

Masumi-compliant agent service with CrewAI integration for the Masumi Network.

Quick Start

# Setup environment
cp .env.example .env
# Add your OpenAI API key to .env

# Install dependencies
uv venv && source .venv/bin/activate
uv pip install -r requirements.txt

# Test the service
uv run python agentic_service.py --input "Hello world" --verbose

# Start API server
uv run python main.py api

Configuration

Required Environment Variables

# OpenAI Configuration
OPENAI_API_KEY=sk-your-openai-api-key-here
OPENAI_MODEL=gpt-4.1-nano  # default: gpt-4.1-nano (cheapest, fastest)

# Masumi Network Configuration
PAYMENT_SERVICE_URL=http://localhost:3001/api/v1
PAYMENT_API_KEY=your-payment-api-key
AGENT_IDENTIFIER=your-agent-identifier
SELLER_VKEY=your-seller-wallet-vkey
NETWORK=Preprod

Customization

The service is designed for easy customization:

• Crew Logic: Edit crew.py for CrewAI agents and tasks
• Agent Config: Modify config/agents.yaml
• Task Config: Modify config/tasks.yaml
• Model Settings: Change OPENAI_MODEL environment variable

Testing

# Test CrewAI integration
uv run python -m pytest test_api.py::TestCrewAIIntegration -v

# Test Masumi compliance
uv run python -m pytest test_api.py::TestMasumiCompliance -v

# Run all tests
uv run python -m pytest test_api.py -v

API Endpoints

Core Endpoints (Masumi Standard)

• POST /start_job - Start new job with payment integration
• GET /status?job_id=<id> - Check job status
• GET /availability - Server availability
• GET /input_schema - Input format schema
• GET /health - Health check

Example Usage

# Start a job
curl -X POST http://localhost:8000/start_job \
  -H "Content-Type: application/json" \
  -d '{"input_data": [{"key": "input_string", "value": "Analyze this text"}]}'

# Check status
curl "http://localhost:8000/status?job_id=your-job-id"

Railway Deployment

1. Deploy Masumi Payment Service

• Provide Blockfrost API key
• Generate public URL after deployment
• Note the URL format: https://your-service.railway.app/api/v1

2. Deploy Agent Service

Railway will automatically detect required environment variables from the Dockerfile. Configure these in the Variables tab:

Required:

• OPENAI_API_KEY - Your OpenAI API key
• PAYMENT_SERVICE_URL - URL from step 1
• PAYMENT_API_KEY - Payment service API key
• NETWORK - Set to Preprod

After agent registration:

• AGENT_IDENTIFIER - From payment service
• SELLER_VKEY - From payment service

3. Register Agent

• Access payment service admin panel
• Register agent using your service URL
• Update agent service with AGENT_IDENTIFIER and SELLER_VKEY

Architecture

This branch demonstrates multi-integration capabilities using git branches:

• main - Minimal reverse echo service (zero dependencies)
• crewai - CrewAI integration (this branch)
• langchain - LangChain integration (planned)
• n8n - n8n workflow integration (planned)

Each branch maintains Masumi compliance while showcasing different AI frameworks.

Development

# Direct service testing
uv run python agentic_service.py --input "test" --verbose

# Local development
uv run python main.py api

# Validate configuration
python get_payment_source_info.py

For detailed Masumi Network documentation, visit docs.masumi.network.