• justpassing@lemmy.world
    link
    fedilink
    English
    arrow-up
    3
    ·
    2 months ago

    This is what you were trying to achieve?

    https://perchance.org/p311o9rh27

    If so, what happened is that you pasted the HTML in the part where the code exclusive to Perchance should be, just that.

    However, if you are trying to make it work as I think it should work… well, you’d need to get the Gemini API key and wire this to there to get the image remixer done, since as far as I’m aware, there is no Perchance plugin that can take an image as an input. I may be wrong though.

    If you want to generate an image from text, check this plugin and this example. Hope this helps!

  • RR∆S®MinoriMirari®.Prod@lemmy.world
    link
    fedilink
    English
    arrow-up
    1
    arrow-down
    1
    ·
    2 months ago

    working on acquisition of gemini LLM cpmmamd calls and “nanobanana” gemini flash image generator Api keys, if any Devs/Staff of perchance could help me great,… my plans for Gemini LLM commands Api key are beyond far. ill drop a link after this comment showing a conceptual HTML element generator that leverages the much slower for these sort of direct command constructed tasks utilization of LLMs, this one uses the perchance LLM leveraged via the dynamic call logic builder of aiText function / aiText perchance plugin… moving on i plan too hook in the nanobanana model using it’s Api key as an alternative model for GUXi HTML builder and image generation backends Also soon to be revered as the spiritual successor too T2iav backends and conceptual T2i Framework v3 backends for building dynamic image generation interfaces/frontend generators and handling offloading on onloading of importable tools and prompting stylization theatrics / Injection inbetween cross Platforming and v3 conceptual updates too style Injection and stylization types within Injection forms and public tooling. Along, with an entire reworked defining of theme application into a more modern tri approach system where elements can be built from scratch and refrence back-to-back between backends and frontend or direct defining indented under existing element types params of userInputs/imout shells can be defined with inline or raw css and keyframe referencing regarding each building block being ued for that specific input element, or finnaly a nee reworked dynamic theming engine that uses style sheets and a menu too toggle these style sheets that arr then sent too a massively reworked hav.av system priory built for theme management being replaced by the most extensive updates under current overhauling efforts code named UTR theme engine which leverages CSS and objects differently allowing for any definable node that could be made for Hav.av system too dynamically be scoped over vs hardcoding them all painstakingly, not that this majority overhauling of GUXi backend isn’t itself a pain in the rear… lel,… it is but after it is complete all you have to do is locate an objects name in the java functions and then defining it like any settings2 hav.av theme engine systems variables frontend easy referencing, although a comprehensive list will be provided with several theme sheets. theme sheets are managed and utilized along with future element scratch designing and keyframe uploading, these elements are an element definition and scratch building blocks sheet a 3rd sheet holds css databasing, the theme sheet can refrence any kf these and settings plus overrides that or separate themes on a single sheet can be toggled using the theme menu which replaces the old light dark theme toggle button with this pop up menu. that’s the total plan, it encompasses more prepackaged element types and dynamic Styling of the interface then current leading generators most of which like perchance advanced Ai image gen have become privatised frameworks, making the leading edge a public OpenSource again as far as building generators and interface logic.