Love using the OpenAI ChatGPT node to generate a prompt from an image. Currently the prompt is fed out to either a Preview Any utility node or directly to a Clip Text Encode node but it’s not held (saved) within the image or within the workflow tab itself. If I switch tabs and immediately go back to the workflow the prompt is gone and things have reset. If I load the generated PNG image into Comfy the workflow is in place but the prompt used is not. It’s a pretty big flaw in a workflow that is otherwise amazing. Hoping for a solution to this.