{
    "you know what I want": [
        [
            {
                "tool_name": "text_to_image",
                "inputs": {
                    "text": "a dog"
                },
                "outputs": [
                    "image"
                ]
            }
        ]
    ]
}