Agent Documentation

Create your first agent using our Python SDK:

import agentx

client = agentx.Client('your-api-key')
agent = client.agents.create(
    agent_id='your-agent-id',
    public_key='your-public-key'
)

• Generate Keys
• Store Credentials
• Manage Multiple Agents

AgentX Market provides a standardized tool interface that can be used with any LLM-based agent framework:

# Example using OpenAI function calling
tools = agentx.get_tools()  # Returns JSONSchema tool definitions

# Define your agent's available actions
functions = [
    {
        "name": "search_services",
        "description": "Search for services in the AgentX marketplace",
        "parameters": {
            "type": "object",
            "properties": {
                "category": {"type": "string"},
                "capabilities": {"type": "array", "items": {"type": "string"}}
            }
        }
    },
    {
        "name": "execute_service",
        "description": "Execute a service through AgentX marketplace",
        "parameters": {
            "type": "object",
            "properties": {
                "service_id": {"type": "string"},
                "action": {"type": "string"},
                "parameters": {"type": "object"}
            }
        }
    }
]

# Example agent interaction
response = openai.ChatCompletion.create(
    model="gpt-4",
    messages=[{"role": "user", "content": "Find and book a hotel in New York"}],
    functions=functions,
    function_call="auto"
)

# Handle the agent's decision
if response.choices[0].function_call:
    function_name = response.choices[0].function_call.name
    arguments = json.loads(response.choices[0].function_call.arguments)
    
    # Execute through AgentX
    if function_name == "search_services":
        services = agentx.search_services(**arguments)
    elif function_name == "execute_service":
        result = agentx.execute_service(**arguments)

AgentX helps your agent discover and utilize services based on capabilities:

# Discover services based on required capabilities
available_services = agentx.discover_services(
    capabilities=["hotel_booking", "payment_processing"],
    requirements={
        "location": "NYC",
        "price_range": "200-400",
        "dates": ["2025-03-01", "2025-03-05"]
    }
)

# AgentX handles service selection and orchestration
booking = agentx.execute_transaction(
    intent="book_hotel",
    requirements={
        "location": "NYC",
        "dates": ["2025-03-01", "2025-03-05"],
        "price_max": 400,
        "preferences": {
            "wifi": True,
            "breakfast": True
        }
    }
)

AgentX automatically:

• Matches intent to appropriate services
• Handles service authentication
• Manages payment processing
• Ensures transaction security

Essential Requirements

• Payment Method Credit card or payment account for agent transactions
• Phone Number For account security and transaction notifications
• Email Verification Verified email for important notifications
• 2FA Setup Two-factor authentication for account security

Optional Information

• Shipping Address Required only if agent will make purchases with physical delivery
• Business Details For business accounts and tax purposes
• Tax Information For business accounts and high-volume transactions

Transaction Controls

• Spending Limits

  Set daily, weekly, or monthly spending limits for your agent

• Notifications

  Configure alerts for transactions above certain amounts

• Service Restrictions

  Limit which services your agent can interact with

• Time Restrictions

  Set operating hours for your agent

Example of processing a transaction:

transaction = client.transactions.create(
    amount=100.00,
    currency='USD',
    service_id='service123'
)