CivArchive
    XUPLX_UnusualPhotoMontageLoRA - - Z IMAGE TURBO -
    Preview 115694312
    Preview 115694313
    Preview 115694793
    Preview 115694687
    Preview 115694314
    Preview 115694780
    Preview 115694315
    Preview 115694799
    Preview 115694881
    Preview 115694785
    Preview 115694722
    Preview 115694311
    Preview 115694740
    Preview 115694448
    Preview 115694948
    Preview 115713046
    Preview 115706168
    Preview 115706179
    Preview 115703832

    UPDATE: Workflow was just published - https://civarchive.com/models/2265145?modelVersionId=2549657

    This is a style LoRA for Z-Image that pushes almost any prompt into a surreal, mid-century photo-montage / cut-paper collage look. It flattens forms into graphic shapes, adds a “printed” texture, and tends to separate foreground figures from clean, poster-like backgrounds.

    What it’s good at / what it’s for

    • Turning normal Z-Image outputs into vintage collage illustrations – think 50s–70s magazine cutouts, postcard art, and book covers.

    • Figurative scenes: standing characters, families, beach scenes, dancers, portraits, etc., arranged like pieces on a page.

    • Still lifes & objects (cups, fruit, cars, lamps, shells, etc.) with bold silhouettes and simple, matte color.

    • Cityscapes, houses, interiors and landscapes stylized into flat, graphic compositions with strong perspective and clean shapes.

    • Fast iterations: strong effect even with low steps (4–20) and LoRA strength around 0.8–1.7, so it’s great for ideation and poster-style pieces.

    • This LoRA affects the model even without a trigger word; use lower weights (~0.6–1.0) if you want a lighter touch.

    Use it when you want the collage look to take over and unify very different prompts into the same visual language.

    How it shouldn’t be used

    • It’s not meant for photorealism. It will fight you if you try to use it for realistic photography, product renders, or ultra-detailed cinematic images.

    • Not ideal for anime, 3D, or painterly oil-painting styles – it tends to override them and drag everything back into collage land.

    • It can slightly distort anatomy and perspective when pushed very hard; it’s a style tool, not a “fix my hands” or “make it realistic” LoRA.

    • Because the effect is strong even without a trigger word, it’s not the right choice if you only want a subtle texture overlay on top of another style LoRA.

    If you need clean realism or a minimal, modern look, leave this LoRA off. If you want your scene to look like a weird, vintage photo-montage pulled from an old art magazine, this is what it’s for.

    DON'T BUY ME A COFFEE!!

    Download ExportGPT - Media & Chat Export Toolkit from the Chrome Web Store. The chrome extension works with ChatGPT and Sora to export assets quickly. I update it often, it has 5 stars, and it's free forever!

    Visit exportgpt.org to see it in action.

    Description

    LORA
    ZImageTurbo

    Details

    Downloads
    126
    Platform
    CivitAI
    Platform Status
    Available
    Created
    12/30/2025
    Updated
    1/9/2026
    Deleted
    -
    Trigger Words:
    xuplx

    Files

    XUPLX_UnusualPhotoMontageLoRA.safetensors