Overview

• Introduction
• Dependencies
• Example
• Final Thoughts

Introduction

AI agents are one of the hottest trends in 2025. There are lot's of improvements, new technologies, libraries in this field. In this post, we'll take a look at the Phidata framework and how it can be used to create an AI agent.

Dependencies

• Python
• Pip
• Gemini

Example

In this example, we'll create a quote generator agent. This agent will take in a subject, author, or keyword from the user, search for relevant quotes, and return a formatted response with a brief explanation of the quote's meaning.

from phi.agent import Agent
from phi.tools.sql import SQLTools
from phi.model.google import Gemini
from phi.tools.duckduckgo import DuckDuckGo
from phi.storage.agent.sqlite import SqlAgentStorage

gemini_model = Gemini(api_key="")

agent = Agent(
    name="quote-generator-agent",
    model=gemini_model,
    instructions=[
        "When the user inputs a subject, author, or keyword, search for relevant quotes.",
        "Provide the quote along with the author's name.",
        "Format the response as: 'quote' - author",
        "Additionally, provide a brief explanation of the quote's meaning."
    ],
    tools=[DuckDuckGo()],
    add_history_to_messages=True,
    num_history_responses=10,
    markdown=True,
    monitoring=True,
)

agent.cli_app(markdown=True)

Example Response

Final Thoughts

In this post, we've covered a simple use case for AI agents. The Phidata framework offers a flexible platform to build advanced assistants, including multi-model agents where multiple agents work together to achieve a goal. With so many exciting developments on the horizon, 2025 is shaping up to be a groundbreaking year for AI agents, operators, and more.

Stay tuned for more updates and happy coding!