Hi, I am Alexander. While I am a developer (Unity, but it is a mess now :D ), my initial AI goal is to help me with writing (not write for me), and do some visual writing (AI-assisted comics, etc) Because of my development background, I see “vibecoding” as a creative pain (it works fast, but rarely exactly as I want, sometimes close enough is good and sometimes not so much) I am dabbling in AI video, but with AI video being even worse slot machine than images, my current approach is - save for DGS SPARK, it will be slower than 5090, but will work and way less electricity. So I hope to go deep into video when I get that goal done. Let's see, eventually, how local images and video can compete with API options. Oh, and I created )vibecoded) some ComfyUI nodes for work with the Google NanoBanana API, but I need to update API endpoints for a month now and can’t get to it for some reason. I hope to be crazier with Comfy when I get any local setup working eventually.

@Nolan Michaels NanoBanana API is paid (with a trick) and does not require local hardware, but I will try to update the Comfy package in few days and then can make a post about it.