

More context, or example could be better to see whats wrong and what the AI is being confused with. Hopefully the 3 characters you have arent 2000 word essays each right ?
As the way i format stuff for the overview, the AI doesnt seem to have issues with identifying who is who, but i also write with way more grounded expectations that with the tiny token window of Perchance. And its not even that different from what you have, only slightly more…meta.

And so on, this can extend to other characters, but again, i dont try to write essays on theology and philosophy with them.







There is no such thing as video generator on Perchance, any generator that claims to do so is either 1) Fake and you got baited 2) Is just a front for another website where it re-routes your prompts, so that means - censored, slow and limited, often all 3 of these.
Same with img2img, Perchance has no such function, anything that claims it can do it is fake or reroutes the requests to another website and to my knowledge there is no img2img website that has the same freedom as Perchance. So any such request will be limited and censored.