qinerous

instead of type = select

write type = visual-select
and you'll understand what I mean

Which generator page you use? Try another generator, There are a lot of generators in Perchance

First the style and background to set the basis for the whole if if the most important thing for you is the style in which the image is drawn. If the most important thing for you is the foreground and the character, then the character description first, and only after its appearance and actions, and the style is at the end. Well, in general, you can probably just mix it all up and scatter it. But it's better to have a structure, since the generator model uses a token system to draw, and the most important thing should written at the very beginning

This is possible if you do not use the text-to-image plugin model from Perchance, but a third-party model that can do this. There may be open source models on github, as well as in huggingface.co . Ask any AI and it will help you with the integration.

idk, I have banned only 1 person on my page in all this time, and I did it without a tag, and his messages were deleted. Well, maybe it doesn't matter with a tag or not :0

official image gen also dont use random tags in banned user list

and in itself plugin site author wrote without tags in example https://perchance.org/comments-plugin

Note that a user's ID is based on their IP address, so you're actually banning a particular IP address. This is usually fine, but sometimes it can result in banning several people - e.g. if a school has a single IP address for all students, then everyone using that school's wifi will have the same user ID. You may want to have a separate comments area with no bannedUsers list for people that are requesting to be unbanned due to e.g. being at the same school as someone else who was previously banned.

and you are using the ban list incorrectly, what you are writing and the tag that is randomly assigned to the user when logging into the site. Its should not be there.

  // DEC0-a445178fe0f9a41e853d
  // 4Q7U-2e3486e5250e20973993
  // 73BN-6ba259ab643205e004c0 (Surfshark EU VPN)
  // DXTI-5eace6d109e009747688
  // I67S-5b53a01a7bdecbd44adc    
  // 9BV8-0674bdbbd00c8bcc3beb

DEC0-a445178fe0f9a41e853d

 thats will be correct
  // a445178fe0f9a41e853d
  // 2e3486e5250e20973993
  // 6ba259ab643205e004c0 (Surfshark EU VPN)
  // 5eace6d109e009747688
  // 5b53a01a7bdecbd44adc    
  // 0674bdbbd00c8bcc3beb

I think it filters by prompt, not by image content. So age deviations don't really matter for the filter here.

@perchance@lemmy.world @VioneT@lemmy.world

thanks for answer

Idk what you mean, but you need just add "adminlair" to right panel where java script codes. In this

let config = {
...
...
...
adminPasswordHash: "----" ,
adminlair: "[MODE]",
...
};

The left panel would work if you used framework plugins , but there use right panel

oh i forgot one line

const textarea = document.querySelector('textarea');
textarea.spellcheck = true;
textarea.autocomplete = "on";
textarea.autocorrect = "on";
textarea.autocapitalize = "on";

tell me later if it worked or not