https://ai.google.dev/gemini-api/docs/computer-use Google released gemini computer use model. It will response action in function_calls. So do i have to use Modular API to work with this model? Is there a better way to do it? Here is offical demo. https://github.com/google/computer-use-preview/blob/main/agent.py

i've actually been looking into this and trying to see exactly how to make the right web request google wants for their API