This example demonstrates how to create an AI agent that generates images from existing images using the Fal AI API.

​
Code

cookbook/agents/multimodal/image_to_image_agent.py
from agno.agent import Agent
from agno.models.openai import OpenAIChat
from agno.tools.fal import FalTools

agent = Agent(
    model=OpenAIChat(id="gpt-5-mini"),
    id="image-to-image",
    name="Image to Image Agent",
    tools=[FalTools()],
    markdown=True,
    instructions=[
        "You have to use the `image_to_image` tool to generate the image.",
        "You are an AI agent that can generate images using the Fal AI API.",
        "You will be given a prompt and an image URL.",
        "You have to return the image URL as provided, don't convert it to markdown or anything else.",
    ],
)

agent.print_response(
    "a cat dressed as a wizard with a background of a mystic forest. Make it look like 'https://fal.media/files/koala/Chls9L2ZnvuipUTEwlnJC.png'",
    stream=True,
)

​
Usage

1

Create a virtual environment

Open the Terminal and create a python virtual environment.
python3 -m venv .venv
source .venv/bin/activate
2

Install libraries

pip install -U agno
3

Run Agent

python cookbook/agents/multimodal/image_to_image_agent.py