Question: Best way to pass an image to Agent with a tool

I'm looking to have a tool do something like this:

def load_image(path: str) -> str:
    return Path(path).read_bytes()

What is the best way to do this?

same question

I usually use two methods:

Use a URL, the agent will take this url as a parameter to the tool.
Set google/adk/agents/ run_comfig.py file
save_input_blobs_as_artifacts: bool = True
Then the user can upload the image in the Adk web, and the agent can take the artifact file name as a parameter to the tool.

Navigation Menu