you can use the vision agent to pass an image URL to have your bot and have GPT analyze the image - I showed this off in a recent livestream 🙂