Dev Documentation

AI Agent Overview

AI Agents: Core Components

When submitting an AI agent to the Data3 Network marketplace, developers must provide key details about their agent to ensure users understand its functionality and can determine if it meets their needs. The components for an AI agent submission are organized as follows:

Section 1: The Information

1. Name and Description

This section is designed to give users a clear and concise understanding of what the AI agent does. The Name should be short, memorable, and directly reflect the agent's functionality. The Description should be simple and straightforward, explaining the core purpose of the agent in a few sentences. A well-crafted name and description ensure users quickly grasp what the agent can do and why they would want to use it.

Example:

• Name: Smart Reminder

• Description: An AI agent that helps you stay on top of your tasks by sending reminders for upcoming events, appointments, or to-do items. It syncs with your calendar and personal preferences, ensuring you never miss an important deadline.

2. Sample Conversations

This section provides examples of how users can interact with the AI agent using natural, conversational commands. The sample conversations illustrate the types of inputs the agent can understand and process. This is crucial for giving users an idea of how intuitive and user-friendly the agent is.

Example:

• User: "Remind me to call Sarah tomorrow at 2 PM."

• AI Agent: "Got it! I'll remind you to call Sarah tomorrow at 2 PM."

• User: "Set a reminder for my meeting at 9 AM tomorrow."

• AI Agent: "Reminder set for your meeting at 9 AM tomorrow."

These examples highlight how users can easily communicate with the agent using plain language.

3. Logo

The Logo is the visual identity of the AI agent, serving as a representation of the agent's brand. As of now, only PNG files are supported for agent logos. The logo should be simple, professional, and distinctive, aligning with the functionality of the agent.

Example:

• The logo for the "Smart Reminder" AI agent could be a small bell or a calendar icon.

4. The Category

Each AI agent is assigned to a specific category to help users find agents that suit their needs. Currently, the following categories are supported:

• Health: AI agents related to healthcare, wellness, fitness, or medical advice.

• Mail: AI agents that assist with email management, scheduling, and communication.

• Social: Agents focused on social media management, interactions, or social networking tasks.

• Finance: Agents designed for financial tracking, budgeting, and accounting.

• Messaging: AI agents for instant messaging, communication automation, and related services.

• Journal: Agents that help users track and manage personal journals, notes, or thoughts.

• Others: For AI agents that do not fit into the above categories.

Example:

• Category for Smart Reminder Agent: Others This category is appropriate because the Smart Reminder agent can work across multiple domains like meetings, tasks, and other personal reminders.

5. Free or Paid

This section indicates whether the AI agent is available for free or requires a paid subscription. If the agent is free, users can start using it immediately without any payment. If the agent is paid, the developer will specify a monthly subscription fee that users will be required to pay for continued access to the agent's features.

Example:

• Free or Paid for Smart Reminder: Free This agent is available at no cost to users who want basic reminder functionality.

6. Minimum Hardware Requirements

The Minimum Hardware Requirements section outlines the basic hardware configuration needed to successfully run or host the AI agent. These requirements ensure that users and node providers understand the computational resources necessary for optimal performance. For AI agents that involve processing large datasets or require real-time computations (like AI-based recommendation systems), higher hardware specifications may be needed.

Example:

• Minimum Hardware for Smart Reminder Agent:

  • CPU: 2 Cores

  • RAM: 2 GB

  • Storage: 500 MB available

  • Network: Stable internet connection

This ensures that users or node providers know what hardware configurations are necessary for the AI agent to run smoothly.

7. The Code Files

To deploy and operate an AI agent, developers must provide the source files for the agent. These files contain the logic and configurations required to run the agent in a containerized environment.

The required files include:

• tools.py: The primary Python script containing the core functions of the AI agent. This file includes the code logic that handles the agent's operations, processes user input, and communicates with external systems (e.g., calendars, email servers).

  Example Content for tools.py:

  def set_reminder(time, message):
      # Code to set the reminder
      pass
  
  def fetch_reminder(time):
      # Code to fetch and display reminders
      pass

• docker-compose.yml: The Docker Compose configuration file used to define the services, networks, and volumes required to run the AI agent in a containerized environment. This file ensures that the agent can be deployed easily on different platforms, including cloud environments and local machines.

  Example Content for docker-compose.yml:

  yamlCopy codeversion: '3.8'
  services:
    reminder-service:
      image: reminder-agent:latest
      container_name: reminder-agent
      ports:
        - "5000:5000"
      volumes:
        - ./data:/app/data
      environment:
        - ENV_VAR=value
      networks:
        - agent-network
  networks:
    agent-network:
      driver: bridge

• Source Files for the Docker Compose Project: The source code files that form the complete AI agent project, including any dependencies, libraries, or helper files necessary for the agent’s functionality.

  Example Content:

  • reminder_agent/

    • app.py

    • requirements.txt

    • config.yaml

    • Dockerfile