CivArchive
    Preview 149435770
    Preview 149435775
    Preview 149435791
    Preview 149435777
    Preview 149435797
    Preview 149435826
    Preview 149435808
    Preview 149435810
    Preview 149435779
    Preview 149435825
    Preview 149435835
    Preview 149435844
    Preview 149435858
    Preview 149435877
    Preview 149435874
    Preview 149435878
    Preview 149435880
    Preview 149435881
    Preview 149435883
    Preview 149435884

    Stop prompting from scratch, explore 11,000+ Prompts for free ๐Ÿ‘‰ PromptDexter.com


    LoRAs for reducing the soft & smooth AI look and adds moreย texture and details to images. It works especially well for close-ups and medium shots where clothing details and skin texture plays an important role.

    My Workflows - https://github.com/vizsumit/comfyui-workflows


    Version Comparison

    Moss_V3: Trained on 3x bigger SFW dataset than previous with no quality compromise, mostly human subjects. It produce much better compositions and colors, also fixes minor anatomy glitches of moss sft. Doesn't lean towards any particular ethnicity so you can generate characters of any ethnicity.

    Moss_V2: Trained on SFW images, more human subjects in dataset and better training strategy. Prefer this for generating pics of human subjects. Not that good with artworks and complex text.

    Moss_V1: Trained on SFW images with wide variety like portraits, lifestyle, fashion, wildlife, sports fitness, architecutre, nature and more.

    Krea_V2_Pro: Trained on professional photography SFW images with wide variety like portraits, lifestyle, fashion, wildlife, sports fitness, architecutre, nature and more. Prefer this for professional looking high-quality images.

    Krea_V1: Trained on SFW and NSFW images so it can handle both. Prefer this for natural looking high-quality images. It also reduces asian face bias.

    KL9_V5_Pro: Trained on professional photography SFW images with wide variety like portraits, lifestyle, fashion, wildlife, sports fitness, architecutre, nature and more.

    KL9_V4: Trained on SFW and NSFW images so it can handle both.

    KL9_V3: Trained on SFW images. It applies very subtle skin texture which looks more realistic, without producing freckles and burnes.

    KL9_V2: Trained on SFW images. It produces more natural looking skin texture than V1, can produce freckles and moles for some time.

    KL9_V1: Over-trained, Don't use this.


    ๐Ÿ–ผ๏ธ Generation Settings

    Weight: 0.5-0.8

    Moss Trigger: no trigger needed

    Krea Trigger: no trigger needed

    KL9B Trigger: This is a high-quality photograph of a (your subject...)


    ๐Ÿ› ๏ธ Editing / Upscaling (Klein9b)

    Weight: 1.0 For V5, 0.5 for later versions

    KL9B Prompt: Make this photo high-quality. Preserve subject's face, expression, figure and pose. Preserve overall composition of this photo.


    Feel free to try it and share results or feedback. ๐Ÿ™‚

    Description

    FAQ

    LORA
    MOSS Image

    Details

    Downloads
    415
    Platform
    CivitAI
    Platform Status
    Available
    Created
    11/18/2026
    Updated
    11/19/2026
    Deleted
    -

    Files

    ultra_real_moss_v3.safetensors

    Mirrors