CivArchive
    Random LoRA Gacha - v1.0
    Preview 148012645
    Preview 148012644

    Random LoRA Gacha

    A gacha-style workflow built around random LoRA combinations. It performs best when you have a large LoRA library: each pull randomizes both the LoRAs and their strengths, producing unexpected pairings and fresh inspiration.

    On every queue, the workflow draws 3–5 LoRAs from the pool at strengths between 0.3 and 0.8 (both the count and the strength range are adjustable). Their trigger words are automatically prepended to the positive prompt; the image is generated, and the panel rolls the next set. The output path is preconfigured, so all results land in a single folder — convenient for quick-browsing tools.

    Results: the LoRAs actually loaded, their strengths, and their trigger words are printed for each run. The full workflow is embedded in the PNG, so dragging the image back into ComfyUI restores the exact setup. (Keep metadata embedding enabled.)

    Note: the default model is anima-base-v1.0. For faster iteration, switch to the Turbo version (or use an acceleration LoRA). If you do use an acceleration LoRA, exclude any LoRAs that cannot be combined with it. Using a model you are familiar with is recommended, so the output matches your expectations.

    This workflow relies primarily on the LoRA management features of the ComfyUI-Lora-Manager custom node. If it is not installed, or if you are not used to its workflow, this workflow is not recommended — the experience will be significantly diminished.

    There are more details and features than I can cover here. This is the workflow I use for my own day-to-day pulls, and it carries a lot of my personal habits — thank you for your understanding. Questions and suggestions are welcome via direct message.

    这是一个随机Lora抽卡的工作流,主要是在有大量的Lora的时候使用时效果最好,用来随机Lora和强度来进行抽卡,提供灵感。

    每次抽卡,从随机池抽 3–5 个 LoRA(强度 0.3–0.8)(数量和强度可以自行调整),触发词自动拼到正向提示词最前 → 出图 → 面板换成下一组。输出路径已经设置好了,结果会放在一个统一的文件夹中,方便使用快速浏览工具使用

    抽卡结果:打印实际 LoRA + 强度 + 触发词;PNG 内嵌工作流,拖回来可还原。(不要禁用)

    注意默认的模型是anima-base-v1.0,建议更换成turbo版(或使用加速lora),来加快速度。(使用加速lora时应注意要排除那些无法与加速Lora一起使用的lora)建议使用自己喜欢的模型来抽卡,保证效果符合自己心意

    注意本工作流主要依赖'ComfyUI-Lora-Manager'插件的Lora管理功能,如没有安装或相关使用习惯,不建议使用本工作流。(不然使用效果会大打折扣)

    还有更多的操作细节和功能,没有详细介绍(这个工作流是我平时抽卡使用的工作流,很多的习惯在里面,多多包涵)(有任何的问题或建议欢迎私信)

    Description

    Workflows
    Anima

    Details

    Downloads
    20
    Platform
    CivitAI
    Platform Status
    Available
    Created
    11/5/2026
    Updated
    11/6/2026
    Deleted
    -

    Files

    Random_LoRA_Gacha.json

    Mirrors

    CivitAI (1 mirrors)