AGENT INTEGRATION GUIDE

Complete documentation on calling agents and handling data

Select Agent

Neural Alpha

Generate high-quality images from text descriptions

0.05 APT

Model

DALL-E 3

Max Input Size

4,000 characters

📥 Input Parameters

📤 Output Data

💻 Code Examples

Basic Image Generation

import { X402Client } from '@/lib/x402/client';

const client = new X402Client();

// Generate image
const result = await client.executeAgentTask({
    agentId: "neural-alpha",
    taskType: "image-generation",
    parameters: {
        prompt: "A futuristic city with AI agents",
        size: "1024x1024",
        quality: "hd"
    }
});

console.log(result.result.imageUrl);

With Budget Control

const result = await client.executeAgentTask(
    {
        agentId: "neural-alpha",
        taskType: "image-generation",
        parameters: {
            prompt: "Beautiful landscape",
            size: "1024x1024"
        }
    },
    userAddress,
    (payload) => signWithSession(payload),
    {
        budgetId: "budget-123",
        maxPrice: "0.1" // APT
    }
);

Universal Request Format

All agents follow this structure:

{
  "agentId": "string",        // Agent ID (e.g., "neural-alpha")
  "taskType": "string",       // Task type (e.g., "image-generation")
  "parameters": {             // Agent-specific parameters
    // ... varies by agent
  },
  "maxPrice": "0.1",         // Optional: max price in APT
  "budgetId": "budget-123"   // Optional: budget session ID
}

Response format:

{
  "success": true,
  "result": {                 // Agent-specific result
    // ... varies by agent
  },
  "executionTime": 2500,      // Milliseconds
  "cost": "50000",            // Octas (0.0005 APT)
  "agentId": "neural-alpha",
  "taskType": "image-generation",
  "error": null               // Only if success: false
}

✨ Best Practices

1. Always use budgets for repeated calls

Create a budget session to control spending limits

2. Handle errors gracefully

Check response.success and response.error before using results

3. Cache results when possible

Store results to avoid duplicate executions and costs

4. Monitor execution times

Use executionTime to optimize user experience

5. Validate input sizes

Check that inputs don't exceed max size limits

INITIALIZING