Hello, first time posting here, so I'm probably going to go overboard with the amount of info, but I'm having an issue where turning on Intelligent image generation (and tested with multi-character both on and off) while self-hosting ComfyUI for image generation causes an issue where my local generator is not called at all to generate an image.
I took a HAR dump and tried to analyze for anything out of the ordinary, but couldn't spot anything.
With intelligentSyncImageV2: true, the client emits no getSyncV2Prompt request at all — no error, no network activity — while narration, TTS segmentation (6–8 segments/beat) and stim dispatch all work normally.
I did notice that
charGenSettings.loras is set to [] for some reason even though the character I selected has a LORA.
This was not the case before the latest update, is anyone else having this issue?
@KairosJS can you assist? Happy to provide any and all details of my local setup.
PS. I've also noticed that when I disable that setting, sometimes the app uses one of the pre-generated images from the character setup gallery, which I'd really prefer it didn't, is there a way to disable this?