I tried couple of things to somehow force the AI to give me full messages, but it seems there is an issue somewhere. I changed the instructions to include an important entry, about chars having no limit on message length. It seems to me it's probably a issue with the chunk loader code, as after getting more than couple paragraphs the text is cut off in the middle of a sentence or word.

hi yo.I have created comment chats through Java script, but I cannot enable emoji packs through them. Either the Chats themselves disappear or nothing happens, and the assistant can't handle it either. Please tell me how

It's not doing parts of the prompt I put in anymore, it's just ignoring it or making it look worse then before.

With helps and efforts, finally, In this testing gen everything seems working and fine testing gen

HOWEVER, when trying to implement the searching function with a real live gallery in gen here, the gallery won't update at all!! (please note how the zoom function works smoothly)

because AI points that the iframe there isn't working

  <div style="width: 100%; height: calc(100% - 65px); overflow: hidden;">
    <div id="galleryContentEl" style="width: 100%; height: 100%; transform: scale(1.0); transform-origin: 0 0;">
      [image(settings.galleryOptions)]
    </div>
  </div>

So it recommended change it from that to this in this gen.

  <div style="width: 100%; height: calc(100% - 65px); overflow: hidden;">
  <div id="galleryContentEl">
    <iframe class="text-to-image-plugin-gallery" src="https://perchance.org/friendly-notification-and-concept-board-animation" style="width:100%; height:100%; border:none; transform-origin: top center;"></iframe>
  </div>
  </div>  

Yes, the searching function works!!. But now the zoom control is a mess; even if that's fixable(harder than it seems); the side panel gallery in the lower right blue button isn't loading at all. (the side gallery is here <div id="chatIframe4" style="width: 100%; height: 100%; overflow: auto; zoom: 0.55; border-radius: 0.25rem;" hidden>[image(settings.galleryOptions)]</div>)

Because I use and only only know vibe coding and now I am complete screwed and cannot vibing anymore. It seems somehow basic but I just cannot get it done because I know nothing.

Please, help!!!

Why isn't there anything in place to control the Chatroom from users who engage in illegal or inappropriate subjects? I understand censorship but I believe there's a fine line between those who keep talking about it and their role in creating or engaging in such activity. Everyday almost hourly there's talk/roleplay/jokes about CP and racism. I'd say this is far from normal behavior. Are these users traceable should anything ever happen? Could child abuse resources be made publicly available on the Perchance website?

i have some interesting side effects here in https://perchance.org/simple-txt2img where adding another row of images refreshes all the images and calling update() only re-generates the last row of images.

how do i add a row of images without re-generating the existing ones? for example if i want to generate until i like the image then create another row to generate more until i like those.

Hi there. Are there any triggers or something similar that can block the operation of $meta? The description and image worked in the site of the last updated sites. But after +-extensively updated the site, $meta was simply ignored. What could be the reason for this? Using meta in js didn't work either. And ai assistant of coding can't handle either.

Hey is there any way to create plugin like the private gallery that saves the images in catbox.moe or perchance.org file or any other cloud storages instead of the device storage?.. My phone storage is almost full.

https://perchance.org/tz4t341n8c#edit In this test generator, in the center, I have a select form which is just a pose list put into options. I have a lot of them in my main generator, and I want to make them a little be more user friendly and interactive, also bypassing language barrier, so I want to add pictures into the options instead( or in addition to words). That's what I did in the poseRadio list. I substituted radio label's background with pictures using CSS styles. I'm planning to have a second list of picture urls for each list, or I could create an array of URLs in the html part. Ideally, it should look like a dropdown with a picture and then the list item name next to it. And when selected, only the item name is stored in the variable without the picture. I'm not sure how to do that yet. Need your help combining select and input radio in some way, thanks!

2025-06-17 1530-1610 GMT time, server gets unreliable, states "Error: image expired" just seconds after creating it, also "Slow down" when adding to galleries. Looks like the server hamster are getting exhausted.

https://perchance.org/ai-text-to-image-generator

just curoius. thanks

Weird question maybe... I'd like to "chain" two generators together in sequence, via a button that passes a bunch of variables from the first one to the second one. Is this doable?

Here's the https://perchance.org/prompt-hunter , a gallery tag searching gen.

    if (formatTagName !== "no filter, currently showing all tags" && dataMatch) {
  
      // Modify the 'bannedPromptPhrases' array
      dataObj.bannedPromptPhrases = [`/^(?!.*(\\b${formatTagName}\\b)).*$/i`];
      localStorage.setItem("lookForThis", formatTagName);

    } else {
      localStorage.removeItem("lookForThis");
    }

    if (formatBadTags !== "no bad words to filter") {
      dataObj.bannedPromptPhrases.push(...formatBadTags);
      let badTagsJoin = formatBadTags.join(', ')
      localStorage.setItem("dontLookForThis", badTagsJoin);
      
    } else {
      localStorage.removeItem("dontLookForThis")
    }

I want to add my own banned list [jail.bannedPromptPhrases] and import jail, added after that as

      if (Array.isArray(jail?.bannedPromptPhrases)) {
        dataObj.bannedPromptPhrases.push(...jail.bannedPromptPhrases);
      }
    }

Now the problem is that's giving unstable results, sometimes it works and sometimes not. I have no idea how or why. In this forked gen https://perchance.org/prompt-hunter#sharedlink=gen%3Aai-furry-generator&galleryname%3Atest3=&tag%3A=&bannedword=: Apple should be banned in PG-13 because it has a PG-13 banned tag "ContentLabel". But sometimes it just won't work!(what's tricky is sometimes it also work) Help!

Hello, I have a problem. I use https://perchance.org/ai-chat, constantly accessing the site from my phone. But in the last week, some problems have been happening, the entire site, not just the chat, can refuse to load and hang either with a simple white screen or with an endless loading. Only rebooting and/or closing and opening the site no less 10 times helped, but usually it was necessary to do this even more times and it somehow connected? But today even this stopped helping, the site does not open at all. When using a VPN, the site opens and seems to work, but then the AI ​​generator itself refuses to work, giving that error verification with an endless connection in the corner of the screen. I'm confused, nothing has changed in my phone or Wi-fi. My main browser is Opera, but when I try to access the site from other browsers, built into the phone and Chrome, the same thing happens, the site does not load. I think it's worth adding that it's from Russia. But I haven't found any information about perchance.org being blocked or something like that.

Hi everyone,

I'm having a very persistent issue with my Perchance generator, and I've tried many troubleshooting steps, but I can't seem to get past it. I'm hoping someone here can take a look.

My Goal: I'm building an art style concept generator. It's a two-generator setup:

my-styles-list: This generator acts purely as a library, holding all my art style lists (General, Dragon Ball, Demon Hunters). my-art-generator: This generator @imports the lists from my-styles-list, uses a selectedCategory dropdown to let the user choose a style category, and then generates a concept based on that choice. The Problem: When I open my Main Art Style Generator, I see:

"My Awesome Art Style Generator" (so the [title] is working). But immediately underneath, it says: "undefined" Below that, it says: "syntax error" Crucially, the "Choose Art Style Category" dropdown menu is NOT appearing on the page. Only the "randomize" button is visible. My Main Art Style Generator's Code:

Perchance Code (Top-Left Editor):

Code snippet

// This is your MAIN Art Style Generator. // It imports lists from your Art Style Library and generates general art style concepts.

// IMPORTANT: Make sure 'my-styles-list' is the EXACT URL slug of your // Art Style Library generator (e.g., 'my-styles-list'). @import:my-styles-list

//

Generator Title

// This list provides the text for the [title] in the HTML template. title = My Awesome Art Style Generator

output Your Art Style Concept: [chooseSelectedStyle]

//

User Input for Style Category

// This creates the dropdown menu on the right-hand side. selectedCategory label = Choose Art Style Category type = select options General Dragon Ball Demon Hunters

//

Conditional Style Generation

// This section uses the user's selection to generate the appropriate style description. chooseSelectedStyle [selectedCategory == "General"] { A [[generalArtStyle]] style, featuring [[generalElements]] and a [[generalColorPalette]] palette, with a [[generalMood]] and [[generalDetails]]. } [selectedCategory == "Dragon Ball"] { A [[dbArtStyle]] style, featuring [[dbElements]] and a [[dbColorPalette]] palette, with a [[dbMood]] and [[dbDetails]]. } [selectedCategory == "Demon Hunters"] { A [[dhArtStyle]] style, featuring [[dhElements]] and a [[dhColorPalette]] palette, with a [[dhMood]] and [[dhDetails]]. } HTML/CSS Code (Bottom-Right Editor):

HTML

What I've Tried So Far (Extensive Troubleshooting):

Library Generator (my-styles-list) is confirmed clean: When I open https://perchance.org/my-styles-list in a private/incognito window, the right-hand side is completely blank, as it should be for a data library. There are no errors on that page itself. Cleared/Reset both generators: For both generators, I've completely deleted all code from both the left (Perchance) and right (HTML/CSS) editors. Used Plain Text Editor intermediary: I copied all the correct code (both Perchance and HTML as shown above) into Notepad/TextEdit first to strip any hidden characters, then pasted it into the blank Perchance editors. Tested in Private/Incognito Mode: All testing has been done in a fresh private window to avoid caching issues. Despite all these steps, the "undefined" and "syntax error" persist, and the dropdown is missing.

My Main Art Style Generator Link: https://perchance.org/my-art-generator

Any insights or help would be greatly appreciated! Thank you

The Perchance dev noted that the new image model "is still training" and "will steadily improve in quality and gain new knowledge over the next few months." Some users found this wording suggests possible self-improvement or ongoing learning. However, in standard practice, most AI models do not learn or evolve after deployment unless manually retrained by developers.

Some questioned whether the model is truly learning over time or whether this statement is simply a placeholder for continued dev-side maintenance.

Goal of the Question: To clarify based on common AI practice:

Is this model capable of continuous or micro-scale learning post-deployment?

If so, can users contribute indirectly (e.g., through use, rating, or feedback) to guide its development?

Or is all improvement strictly the result of developer-side retraining and version updates?

Understanding this helps the community know whether their activity matters in shaping the model—or if they should wait for official updates. It also opens the door to potential participatory development, if such feedback loops are supported.

Hi folks,

I want to be able to generate an image and manipulate it on the page in relation to other elements.

I know that when using the image generator we can access the dataUrl from the fulfilled Promise to get the raw image. eg.

let window.generated_images = []
...
image = {import:text-to-image-plugin}

let promptData = {
      prompt: "blah",
      negativePrompt: "Not blah",
      resolution: "512x768",
      style: "margin:0.25rem",
    };

let generator = image(promptData);

generator.then(x=>function(x){
  window.generated_images.push(x.dataUrl);
});
...
// Add all generated_images to the page by creating and adding <img> elements.
// Do some cool stuff like automatically detecting and appending images to sections depending on associated conditions.

But by doing this we lose access to a number of features which a packaged with the iframe. eg.

1. Saving to Gallery
2. Inspecting the prompt
3. "Chat With This Character" button etc.

When attempting to access the elements within the relevant iframe I get with an XSS error.

Is there anyway for me to manipulate the iframe contained image elements after they've been generated? Or to add the lost functionality back to the dataUrl images?

Ok,I get that this thing has been "reset" and no longer recognises famous faces (I didn't realise a law had been passed to not use them btw)...So,just a simple question...is this something that in time will get back to where we were a few weeks ago on the previous version,where it would recognise famous faces instantly or is there now something built into this model that prohibits it from doing that?...trying to recreate famous people on this new version (if that's your thing) via prompts is fairly soul-destroying...just seem to keep getting the exact same face.

do you guys know any specific artists the generator recognize? old one used to have some anime styles by default but idk about Flux

Need to know why past week I been getting random issues with ad blocker or firefox causing it to fail to generate, just sits & spins. I disable them when on perchance but it sty whines about it & causes problems. either an update pissed off firefox or one of my ad blockers & would like the admin to look into this soon as ur able if you please.

there are times it works fine with them ad blockers on/disabled but then suddenly specially if u have browser open/background it freaks out & refuses to work, despite clearing cache & refreshing. I use ad blockers for streaming & do not like perm disabling to come here, it should be good enough to disable on perchance.org since disable does not block your ad's.

Hey, i'm quite new to perchance so i'm not really familiar with it yet. sorry if my question is stupid. i was wondering, when i create a character, if there is a way to save it and use it as a reference. as a small example: i would like to create graphics for a storyline or a comic. but for that i need a character that looks the same in several situations. is that somehow possible?

I'm working on a test generator for gallery prompt searching here: https://perchance.org/gallery-search-forking-rudbo

It’s forked from prompt-hunter, created by the author of AI Art Generator (Advanced).

The issue arises when I try to implement this into another test generator: https://perchance.org/friendly-notification-and-concept-board-animation In this version, the search button doesn’t update the gallery as expected.

It seems like the gallery either needs to be enforced via iframe or replaced with a local component—but neither approach has worked so far.

Strangely, if I click the toggle theme button, the search button suddenly works once. So I tried adding the following logic to the search button:

onclick="changeAll(tagInputEl.value); update(panelCtn);"

and then also tried:

onclick="changeAll(tagInputEl.value); update(panelCtn); update(galleryContentEl);"

Still no effect.

Any suggestions or insights would be greatly appreciated!

I do not like the new model doesn’t know celebrity faces also. It knows will Smith but not Jack Nicholson, not anyone like the old model knew every celebrity almost. Please put old model back 😭