Skip to main content

Claude MCP Demo

This demo showcases how Claude can use PayStabl to automatically handle payments for premium APIs through the Model Context Protocol (MCP) server integration.

Demo Overview​

In this example, Claude will:

  1. Receive a request for premium data analysis
  2. Automatically pay for required API access
  3. Process the data and return results
  4. Handle payment receipts and error scenarios

Setup Instructions​

1. Install PayStabl MCP Server​

npm install -g @paystabl/mcp-server

2. Configure Claude Desktop​

Add PayStabl to your Claude Desktop configuration:

{
  "mcpServers": {
    "paystabl": {
      "command": "paystabl-mcp-server",
      "args": ["--agent-id", "claude-demo"],
      "env": {
        "PAYSTABL_API_KEY": "your-api-key-here",
        "PAYSTABL_NETWORK": "base-sepolia"
      }
    }
  }
}

3. Fund Your Agent Wallet​

Get testnet funds for demonstration:

# Request Base Sepolia ETH
curl -X POST https://faucet.quicknode.com/base/sepolia \
  -H "Content-Type: application/json" \
  -d '{"address": "YOUR_WALLET_ADDRESS"}'

Demo Scenarios​

Scenario 1: Market Data Analysis​

User Request:

"Get the latest crypto market data and analyze the trends for Bitcoin and Ethereum"

Claude's Response:

I'll get the latest crypto market data for you and analyze the trends.

First, let me access the premium market data API...

[Using pay_for_api tool internally]
✅ Payment processed: $2.50 to CryptoMarketAPI
Transaction: 0x1234abcd...

Now analyzing the data:

**Bitcoin (BTC)**
- Current Price: $43,250
- 24h Change: +2.3% 
- Volume: $28.5B
- Trend: Bullish momentum with strong support at $42K

**Ethereum (ETH)**  
- Current Price: $2,580
- 24h Change: +1.8%
- Volume: $15.2B
- Trend: Steady growth following BTC, resistance at $2.6K

**Analysis:**
The market shows positive sentiment with both major cryptocurrencies gaining. Bitcoin's break above $43K suggests continued upward momentum, while Ethereum is consolidating before a potential breakout.

Payment Summary: $2.50 for premium market data access

Scenario 2: Document Processing​

User Request:

"Analyze this contract PDF and extract key terms and potential risks"

Claude's Response:

I'll analyze your contract PDF for key terms and risks.

[Uploading document to premium analysis service...]

[Using pay_for_api tool]
✅ Payment processed: $5.00 to LegalAnalysisAPI
Transaction: 0x5678efgh...

**Key Terms Identified:**
- Contract Duration: 24 months
- Payment Terms: Net 30 days
- Termination Clause: 60-day notice required
- Liability Cap: $100,000
- Intellectual Property: All rights retained by creator

**Risk Assessment:**
🔴 **High Risk:** Unlimited liability clause in section 4.2
🟡 **Medium Risk:** Vague delivery timelines 
🟢 **Low Risk:** Standard termination provisions

**Recommendations:**
1. Negotiate liability cap in section 4.2
2. Define specific delivery milestones
3. Add force majeure clause

Payment Summary: $5.00 for legal document analysis

Scenario 3: Agent Collaboration​

User Request:

"I need a comprehensive research report on AI agent frameworks. Can you coordinate with specialist agents?"

Claude's Response:

I'll coordinate with specialist research agents to create a comprehensive report.

**Phase 1: Research Coordination**

[Using pay_agent tool]
✅ Paid Research Agent: $8.00
Purpose: AI framework landscape analysis
Transaction: 0x9abc1234...

[Using pay_agent tool]  
✅ Paid Technical Writer: $12.00
Purpose: Report compilation and formatting
Transaction: 0xdef56789...

**Phase 2: Data Collection**

[Using pay_for_api tool]
✅ Paid Academic Database: $6.00
Purpose: Recent research papers and studies
Transaction: 0x1111aaaa...

**Research Report: AI Agent Frameworks 2024**

**Executive Summary:**
The AI agent framework landscape has evolved rapidly, with 3 major categories emerging...

[Full detailed report with citations, comparisons, and recommendations]

**Payment Summary:**
- Research Agent: $8.00
- Technical Writer: $12.00  
- Academic Database: $6.00
- **Total: $26.00**

All payments completed successfully with full audit trail.

Interactive Demo Scripts​

Basic Payment Demo​

Create a simple script to test Claude's payment capabilities:

// demo-script.js
const readline = require('readline');

const rl = readline.createInterface({
  input: process.stdin,
  output: process.stdout
});

const demoScenarios = {
  weather: {
    prompt: "Get premium weather forecast for San Francisco with detailed hourly breakdown",
    expectedCost: "$1.50",
    api: "WeatherPro API"
  },
  sentiment: {
    prompt: "Analyze the sentiment of this text: 'PayStabl is revolutionizing how AI agents handle payments'",
    expectedCost: "$0.75",
    api: "SentimentAnalysis API"  
  },
  translation: {
    prompt: "Translate this technical document from English to Japanese with domain expertise",
    expectedCost: "$3.25",
    api: "TechnicalTranslation API"
  }
};

console.log("🤖 Claude PayStabl Demo");
console.log("========================");
console.log("\nAvailable demo scenarios:");

Object.keys(demoScenarios).forEach((key, index) => {
  const scenario = demoScenarios[key];
  console.log(`${index + 1}. ${key.toUpperCase()}: ${scenario.expectedCost} - ${scenario.api}`);
});

rl.question('\nSelect a scenario (1-3) or type a custom request: ', (answer) => {
  const scenarioKeys = Object.keys(demoScenarios);
  const selectedIndex = parseInt(answer) - 1;
  
  if (selectedIndex >= 0 && selectedIndex < scenarioKeys.length) {
    const scenario = demoScenarios[scenarioKeys[selectedIndex]];
    console.log(`\n📋 Demo Scenario: ${scenario.prompt}`);
    console.log(`💰 Expected Cost: ${scenario.expectedCost}`);
    console.log(`🔗 API: ${scenario.api}`);
    console.log("\n✨ Now ask Claude: \"" + scenario.prompt + "\"");
  } else {
    console.log(`\n📋 Custom Request: ${answer}`);
    console.log("\n✨ Ask Claude this question and watch PayStabl handle payments automatically!");
  }
  
  rl.close();
});

Run the demo:

node demo-script.js

Payment Monitoring Script​

Monitor Claude's payment activity in real-time:

// payment-monitor.js
const { PayStablAgent } = require('@paystabl/sdk');

class ClaudePaymentMonitor {
  constructor() {
    this.agent = new PayStablAgent({
      agentId: "claude-demo",
      network: "base-sepolia"
    });
    
    this.startTime = Date.now();
    this.totalSpent = 0;
    this.transactionCount = 0;
  }
  
  async startMonitoring() {
    console.log("🔍 Monitoring Claude's PayStabl Activity");
    console.log("======================================");
    
    // Display initial balance
    const balance = await this.agent.getBalance();
    console.log(`💰 Starting Balance: ${balance.eth} ETH`);
    console.log(`🆔 Agent ID: claude-demo`);
    console.log(`🌐 Network: Base Sepolia`);
    console.log("");
    
    // Monitor transactions every 10 seconds
    setInterval(async () => {
      await this.checkRecentActivity();
    }, 10000);
    
    // Display summary every minute
    setInterval(() => {
      this.displaySummary();
    }, 60000);
  }
  
  async checkRecentActivity() {
    try {
      const recent = await this.agent.getTransactions({
        limit: 5,
        since: new Date(Date.now() - 15000) // Last 15 seconds
      });
      
      recent.forEach(tx => {
        if (!this.processedTxs.has(tx.txHash)) {
          this.processedTxs.add(tx.txHash);
          this.logTransaction(tx);
          
          if (tx.status === 'completed') {
            this.totalSpent += parseFloat(tx.amount);
            this.transactionCount++;
          }
        }
      });
    } catch (error) {
      console.error("Error checking activity:", error.message);
    }
  }
  
  logTransaction(tx) {
    const timestamp = new Date().toLocaleTimeString();
    const status = tx.status === 'completed' ? '✅' : 
                  tx.status === 'failed' ? '❌' : '⏳';
    
    console.log(`[${timestamp}] ${status} ${tx.type.toUpperCase()}`);
    console.log(`  Amount: $${tx.amount}`);
    console.log(`  Purpose: ${tx.purpose || 'N/A'}`);
    console.log(`  Hash: ${tx.txHash.substring(0, 10)}...`);
    console.log("");
  }
  
  displaySummary() {
    const runtime = Math.floor((Date.now() - this.startTime) / 1000 / 60);
    
    console.log("📊 PAYMENT SUMMARY");
    console.log("==================");
    console.log(`⏱️  Runtime: ${runtime} minutes`);
    console.log(`💸 Total Spent: $${this.totalSpent.toFixed(2)}`);
    console.log(`📈 Transactions: ${this.transactionCount}`);
    console.log(`💰 Avg per Transaction: $${(this.totalSpent / Math.max(this.transactionCount, 1)).toFixed(2)}`);
    console.log("");
  }
  
  processedTxs = new Set();
}

// Start monitoring
const monitor = new ClaudePaymentMonitor();
monitor.startMonitoring().catch(console.error);

// Graceful shutdown
process.on('SIGINT', () => {
  console.log("\n👋 Stopping payment monitor...");
  process.exit(0);
});

Run the monitor:

node payment-monitor.js

Error Scenarios Demo​

Insufficient Funds​

User Request:

"Analyze this large dataset with the most expensive AI model available"

Claude's Response:

I'll analyze your dataset with our premium AI model.

[Attempting payment...]

❌ Payment failed: Insufficient funds in agent wallet

Current balance: $2.15
Required amount: $15.00

I need additional funding to access the premium AI model. Here are your options:

1. **Add funds to agent wallet**
   - Visit: https://faucet.quicknode.com/base/sepolia
   - Request testnet ETH for demonstration

2. **Use standard analysis** (Available for $2.00)
   - Slightly less detailed but still comprehensive
   - Would you like me to proceed with this option?

3. **Request funding approval**
   - I can notify administrators for funding approval

What would you prefer?

Service Unavailable​

User Request:

"Get real-time stock prices for AAPL, GOOGL, and TSLA"

Claude's Response:

I'll get the latest stock prices for you.

[Attempting to access MarketData API...]

⚠️ Primary service unavailable (HTTP 503)

Let me try alternative data sources...

[Using pay_for_api with backup service]
✅ Payment processed: $1.75 to AlternativeMarketAPI
Transaction: 0xbackup123...

**Stock Prices** (via backup service):
- AAPL: $185.43 (+1.2%)
- GOOGL: $142.17 (+0.8%) 
- TSLA: $248.91 (+2.1%)

*Note: Used backup service due to primary API maintenance*
Payment: $1.75 (vs $1.25 for primary service)

Advanced Demo Features​

Multi-Agent Workflow​

Demonstrate complex agent collaboration:

// multi-agent-demo.js
const scenarios = [
  {
    name: "Research & Analysis Pipeline",
    description: "Claude coordinates multiple specialist agents",
    agents: [
      { name: "Data Collector", cost: "$3.00", task: "Gather research data" },
      { name: "Statistical Analyst", cost: "$5.00", task: "Perform statistical analysis" },
      { name: "Report Writer", cost: "$4.00", task: "Generate final report" }
    ],
    totalCost: "$12.00",
    prompt: "Create a market analysis report for the renewable energy sector"
  },
  {
    name: "Content Creation Workflow", 
    description: "Multi-step content generation with specialist agents",
    agents: [
      { name: "Research Agent", cost: "$2.50", task: "Topic research" },
      { name: "Content Writer", cost: "$6.00", task: "Draft creation" },
      { name: "Editor Agent", cost: "$3.50", task: "Content refinement" },
      { name: "SEO Optimizer", cost: "$2.00", task: "SEO optimization" }
    ],
    totalCost: "$14.00",
    prompt: "Create an SEO-optimized blog post about AI agent payment systems"
  }
];

function displayScenario(scenario) {
  console.log(`\n🎬 ${scenario.name}`);
  console.log("=".repeat(scenario.name.length + 3));
  console.log(scenario.description);
  console.log(`💰 Total Cost: ${scenario.totalCost}`);
  console.log("\n👥 Agent Workflow:");
  
  scenario.agents.forEach((agent, index) => {
    console.log(`  ${index + 1}. ${agent.name} (${agent.cost}) - ${agent.task}`);
  });
  
  console.log(`\n📝 Prompt: "${scenario.prompt}"`);
  console.log("\n" + "=".repeat(50));
}

console.log("🤖 Multi-Agent PayStabl Demo");
console.log("Multi-agent workflows with automatic payment handling\n");

scenarios.forEach(displayScenario);

console.log("\n✨ Try any of these prompts with Claude to see PayStabl");
console.log("   coordinate payments across multiple agents automatically!");

Cost Optimization Demo​

Show how PayStabl optimizes costs:

// cost-optimization-demo.js
const costScenarios = [
  {
    task: "Text summarization",
    options: [
      { service: "Premium AI", cost: "$2.50", quality: "Excellent" },
      { service: "Standard AI", cost: "$1.00", quality: "Good" },
      { service: "Basic AI", cost: "$0.25", quality: "Adequate" }
    ],
    recommendation: "Standard AI - best value for quality"
  },
  {
    task: "Language translation", 
    options: [
      { service: "Human Expert", cost: "$15.00", quality: "Perfect" },
      { service: "Advanced AI", cost: "$3.00", quality: "Excellent" },
      { service: "Basic Translation", cost: "$0.50", quality: "Fair" }
    ],
    recommendation: "Advanced AI - professional quality at low cost"
  }
];

function demonstrateCostOptimization() {
  console.log("💡 PayStabl Cost Optimization");
  console.log("==============================");
  
  costScenarios.forEach(scenario => {
    console.log(`\n📋 Task: ${scenario.task}`);
    console.log("Options available:");
    
    scenario.options.forEach((option, index) => {
      const selected = option.service.includes(scenario.recommendation.split(' ')[0]) ? "👍" : "  ";
      console.log(`  ${selected} ${option.service}: ${option.cost} (${option.quality})`);
    });
    
    console.log(`🎯 Recommended: ${scenario.recommendation}`);
  });
  
  console.log("\n✨ PayStabl automatically selects the best value option");
  console.log("   based on your budget and quality requirements!");
}

demonstrateCostOptimization();

Demo Best Practices​

1. Preparation Checklist​

  • MCP server installed and configured
  • Agent wallet funded with testnet ETH
  • Claude Desktop restarted after configuration
  • Payment monitoring script ready
  • Demo scenarios planned

2. Demonstration Flow​

  1. Introduction (2 minutes)

    • Explain PayStabl concept
    • Show agent wallet balance
    • Start payment monitor

  2. Simple Payment Demo (3 minutes)

    • Basic API payment request
    • Show automatic x402 handling
    • Display payment receipt

  3. Complex Workflow (5 minutes)

    • Multi-agent collaboration
    • Real-time payment tracking
    • Cost breakdown explanation

  4. Error Handling (2 minutes)

    • Insufficient funds scenario
    • Service unavailability handling
    • Recovery strategies

  5. Q&A and Discussion (3 minutes)

    • Technical questions
    • Use case discussion
    • Integration possibilities

3. Common Demo Issues​

Issue: Claude doesn't show payment tools

  • Solution: Restart Claude Desktop after MCP configuration

Issue: Payment failures during demo

  • Solution: Check testnet fund balance before starting

Issue: Slow transaction confirmations

  • Solution: Use Base Sepolia (faster than Ethereum Sepolia)

Issue: MCP server connection errors

  • Solution: Verify API key and network settings

Demo Variations​

Technical Audience Demo​

Focus on:

  • x402 protocol details
  • Smart contract interactions
  • Security and policy features
  • Integration architecture

Business Audience Demo​

Focus on:

  • Cost savings and efficiency
  • Autonomous operation benefits
  • Use case scenarios
  • ROI calculations

Developer Audience Demo​

Focus on:

  • SDK integration examples
  • Custom framework implementations
  • Testing and debugging tools
  • Production deployment strategies

Support​

For demo support and questions: