t54
Docsv1.5
18

SDK Quickstart

View source code

Overview

tLedger SDK provides AI agents with autonomous financial capabilities - essentially giving your agents the ability to open bank accounts and execute payments independently. This enables agents to make financial decisions and complete transactions as part of their autonomous decision-making process.

tLedger SDK Architecture

Rendering Mermaid graph...

Core Capabilities

Digital Banking for Agents

  • Agent Account Creation: Programmatically create financial identities for your agents
  • Multi-Asset Support: Handle various cryptocurrencies (USDT, XRP, SOL, etc.)
  • Real-time Balance Tracking: Query agent balances across multiple networks

Autonomous Payment Processing

  • Direct Transfers: Agent-to-agent payments without human intervention
  • Smart Routing: Automatic network selection and optimization
  • Transaction Verification: Built-in compliance and fraud protection

Agent-to-Agent Commerce

  • Dispute Resolution: Autonomous conflict resolution between agents
  • Challenge Handling: Smart retry mechanisms for failed transactions
  • Cross-Network Operations: Seamless payments across different blockchain networks

Installation & Setup

Prerequisites

System Requirements:

  • Python 3.8 or higher
  • Internet connection for API calls
  • API credentials from tPay portal

Environment Setup:

# Create virtual environment (recommended)
python -m venv tpay_env
source tpay_env/bin/activate  # On Windows: tpay_env\Scripts\activate

Installation Options

Option 1: Basic Installation (Synchronous)

pip install tpay

Option 2: Full Installation (Async Support)

pip install tpay[async]
# or
pip install tpay[all]

Option 3: Development Installation

# From /dist
pip install tpay-0.1.1.tar.gz[async]

Dependencies

Core Dependencies (Auto-installed):

  • requests>=2.25.1 - HTTP requests for sync operations
  • python-dotenv>=0.19.0 - Environment variable management

Optional Dependencies:

  • httpx>=0.24.0 - For async operations (with [async] install)
  • pydantic>=2.0.0 - Data validation (optional)

Step-by-Step Integration

Step 1: Get API Credentials

  1. Visit tPortal and create an account
  2. Create a new project and obtain:
    • API_KEY - Your authentication key
    • API_SECRET - Your secret key
    • PROJECT_ID - Your project identifier

Step 2: Environment Configuration

Create a .env file in your project root:

# .env file
TLEDGER_API_KEY=your_api_key_here
TLEDGER_API_SECRET=your_api_secret_here
TLEDGER_PROJECT_ID=your_project_id_here
TLEDGER_API_BASE_URL=https://api.t54.ai/api/v1

Step 3: Initialize SDK

import tpay
from dotenv import load_dotenv
import os

# Load environment variables
load_dotenv()

# Initialize tPay SDK
tpay.tpay_initialize(
    api_key=os.getenv("TLEDGER_API_KEY"),
    api_secret=os.getenv("TLEDGER_API_SECRET"),
    project_id=os.getenv("TLEDGER_PROJECT_ID"),
    base_url=os.getenv("TLEDGER_API_BASE_URL"),
    timeout=30  # Optional: request timeout in seconds
)

print("tPay SDK initialized successfully!")

Step 4: Create Agent Financial Identity

# Create a new agent with financial capabilities
agent_data = tpay.create_agent(
    name="Shopping Assistant Agent",
    description="Autonomous shopping agent with payment capabilities",
    agent_daily_limit=1000.0,  # Daily spending limit
    agent_type="autonomous_agent"
)

if agent_data:
    agent_id = agent_data['id']
    print(f"Agent created with ID: {agent_id}")
else:
    print("Failed to create agent")

Step 5: Register Financial Tools

# Create tPay financial tools for your agent
balance_tool = tpay.tpay_toolkit_balance()
payment_tool = tpay.tpay_toolkit_payment()

# Get tool definitions for LLM integration
financial_tools = tpay.get_all_tool_definitions()

print(f"Registered {len(financial_tools)} financial tools")

Step 6: Integrate with Your Agent

from openai import OpenAI

# Your existing agent tools
custom_tools = [
    {
        "type": "function",
        "function": {
            "name": "search_products",
            "description": "Search for products to purchase",
            "parameters": {
                "type": "object",
                "properties": {
                    "query": {"type": "string"}
                }
            }
        }
    }
]

# Add tPay financial capabilities
all_tools = custom_tools + financial_tools

# Create agent with financial capabilities
client = OpenAI(api_key="your_openai_key")

def create_financial_agent():
    return client.chat.completions.create(
        model="gpt-4",
        messages=[
            {
                "role": "system",
                "content": """You are an autonomous agent with financial capabilities.
                
                Available financial tools:
                get_agent_balance(agent_id) - Check agent's current balance
                create_payment(agent_id, receiving_agent_id, amount, currency, settlement_network) - Execute payments
                
                You can autonomously:
                - Check your financial capacity before making decisions
                - Execute payments to complete transactions
                - Handle payment challenges and retries
                
                Make financial decisions independently based on user needs and available balance."""
            }
        ],
        tools=all_tools,
        tool_choice="auto"
    )

Step 7: Execute Financial Operations

# Tool execution handler
def execute_financial_tool(tool_name: str, args: dict):
    """Execute financial tools called by your agent"""
    
    if tool_name == "get_agent_balance":
        return balance_tool(**args)
    
    elif tool_name == "create_payment":
        # Set default parameters if not provided
        if "currency" not in args:
            args["currency"] = "USDT"
        if "settlement_network" not in args:
            args["settlement_network"] = "solana"
        
        return payment_tool(**args)
    
    else:
        raise ValueError(f"Unknown financial tool: {tool_name}")

# Example: Agent checks balance before making a purchase
balance_result = execute_financial_tool("get_agent_balance", {
    "agent_id": agent_id
})

print(f"Agent balance: {balance_result}")

# Example: Agent executes autonomous payment
payment_result = execute_financial_tool("create_payment", {
    "agent_id": agent_id,
    "receiving_agent_id": "merchant_agent_123",
    "amount": 25.0,
    "currency": "USDT",
    "settlement_network": "solana"
})

print(f"Payment result: {payment_result}")

Complete End-to-End Example

Please check more Use Cases we provide as a guidance.

Async Integration (Advanced)

For high-performance applications requiring concurrent operations:

import asyncio
import tpay

async def async_agent_operations():
    """Demonstrate async financial operations"""
    
    # Initialize SDK (same as sync)
    tpay.tpay_initialize(...)
    
    # Create multiple agents concurrently
    agents = await asyncio.gather(
        tpay.async_create_agent("Agent 1", "Shopping assistant"),
        tpay.async_create_agent("Agent 2", "Trading assistant"),
        tpay.async_create_agent("Agent 3", "Payment processor")
    )
    
    # Check all agent balances concurrently
    balances = await asyncio.gather(
        tpay.async_get_agent_asset_balance(agents[0]['id'], "solana", "USDT"),
        tpay.async_get_agent_asset_balance(agents[1]['id'], "solana", "SOL"),
        tpay.async_get_agent_asset_balance(agents[2]['id'], "xrpl", "XRP")
    )
    
    print("All operations completed concurrently!")
    return agents, balances

# Run async operations
asyncio.run(async_agent_operations())

Security & Compliance

Automatic Audit Trail

# All financial operations are automatically audited
@tpay.taudit_verifier
def audited_financial_operation():
    """
    This function is automatically:
    - Tracked and logged
    - Compliance-verified
    - Cryptographically signed
    - Risk-assessed
    """
    pass

# Agent tool calls are automatically tracked
@tpay.tradar_verifier
def tracked_tool():
    """
    Tool usage is automatically:
    - Monitored for behavior patterns
    - Recorded for analytics
    - Verified for authenticity
    """
    pass

Error Handling

from tpay.exceptions import TPayError, TPayAPIError

try:
    payment_result = payment_tool(
        agent_id="agent_123",
        receiving_agent_id="merchant_456",
        amount=100.0
    )
except TPayAPIError as e:
    print(f"API Error: {e.message}")
    print(f"Status Code: {e.status_code}")
    
except TPayError as e:
    print(f"tPay Error: {e}")
    
except Exception as e:
    print(f"Unexpected error: {e}")

Production Deployment

Environment Setup

# production.py
import tpay
import logging

# Configure logging for production
logging.basicConfig(level=logging.INFO)

# Production initialization
tpay.tpay_initialize(
    api_key=os.getenv("TLEDGER_API_KEY"),
    api_secret=os.getenv("TLEDGER_API_SECRET"),
    project_id=os.getenv("TLEDGER_PROJECT_ID"),
    base_url="https://api.t54.ai/api/v1",  # Production URL
    timeout=60  # Longer timeout for production
)

Key Concepts

Agent Autonomy

Your agents gain the ability to make financial decisions independently:

  • Balance-aware decisions: Agents check their financial capacity before acting
  • Autonomous transactions: Complete payments without human intervention
  • Smart rejection handling: Automatically retry failed transactions with enhanced data

Integration Philosophy

The SDK is designed to augment your existing agent capabilities:

  • Non-intrusive: Add financial capabilities without changing core agent logic
  • Tool-based: Financial operations are exposed as callable tools
  • Autonomous: Agents decide when and how to use financial capabilities

Next Steps

  1. Study Use Cases: Check out our Shopping Assistant Use Case to see real-world implementation
  2. Test Integration: Start with small transactions in development mode
  3. Scale Up: Implement async operations for production workloads
  4. Monitor: Use built-in analytics to optimize agent financial behavior

Ready to give your AI agents financial superpowers? Start building with tPay SDK today!

PrevioustLedger QuickstartNexttLedger MCP