Skip to content
WorkflowVerifiedExternal
Open source

Anime Fugi Creator

as the Angelssword adventurers creator has a lot of tools for removing the green screen and looping.

FamilyQwen-ImageVersionRelease 3099495 — animeFugiCreator.jsonEstimated VRAMEstimate unavailableCgrade

Runtime profile

Access
External workflow
Status
Verified
Version
animeFugiCreator.json
Source
Civitai
Updated
Oct 18, 2026

Source description

I recommend looking at https://github.com/OozeClues/angelssword-adventurers-creator

as the Angelssword adventurers creator has a lot of tools for removing the green screen and looping.

This is a workflow to create animated Fugi's for Discord. Settings are limited so it fits in the 120 frame limit. This workflow probably requires ~20GB of Vram.

Nodes:

KJNodes

WanVideoWrapper

VideoHelperSuite

CorridorKey

note: I think the corridorKey Node is busted but I got it to work by just asking Hermes agent.

Step: 0 install and set up the 5 gazillion models as the is 3 workflow stitched together

Step 1: Insert character turn around

Step 2: use prompt to create a portrait with a green background. (red may also work but I have not tested it)

Step 3: prompt the movement you want for the gif. talking, idle, muted, deafened. (for anime style character I recommend prompting for 3d or CG style)

Step 3.5: Commit AI Gamba

Step 4: profit

Requirements

Models

12

Node packs

3

AI-generated commentary

AI-generated explanation based on source and configuration details. Suggestions are clearly labeled.

Purpose

  • Anime Fugi Creator is a workflow for creating animated Fugi GIFs for Discord: give it a character turn-around, a portrait prompt, and a movement prompt, and it can produce an animated GIF.

Flow

  • Insert a character turn-around, use a prompt to create a portrait with a green background, then prompt movement such as talking, idle, muted, or deafened.

Inputs and outputs

  • Inputs are a character turn-around, a portrait prompt, and a movement prompt.

  • The output is an animated Fugi GIF for Discord.

Setup

  • Plan to set up the listed model files and node packs before use.

    Suggestion · not verified

  • The workflow probably requires about 20 GB of VRAM.

Dependencies

  • The listed node packs are comfyui-kjnodes, comfyui-videohelpersuite, and ComfyUI-WanVideoWrapper.

  • Listed model files include BiRefNet_Dynamic.safetensors, birefnet.safetensors, CorridorKey_v1.0.pth, qwen_2.5_vl_7b_fp8_scaled.safetensors, qwen_image_edit_2509_fp8_e4m3fn.safetensors, qwen_image_edit_2511_int8_convrot.safetensors, and qwen_image_vae.safetensors.

  • Additional listed model files are Qwen-Image-Edit-2509-Lightning-4steps-V1.0-bf16.safetensors, umt5-xxl-enc-bf16.safetensors, wan_2.1_vae.safetensors, Wan2_2-I2V-A14B-HIGH_fp8_e5m2_scaled_KJ.safetensors, Wan2_2-I2V-A14B-LOW_fp8_e5m2_scaled_KJ.safetensors, wan2.2_t2v_A14b_high‌_‌،

Configuration

  • Settings are limited to 120 frames.

  • For anime-style characters, try a 3D or CG-style movement prompt.

    Suggestion · not verified

  • Use a green background for the portrait; red is mentioned as an alternative, but it has not been tested.

    Suggestion · not verified

Cautions

  • The CorridorKey node may not work, so verify it in your setup before relying on it.

    Suggestion · not verified

Does this need to be edited?

Sign in to send an edit request.

Sources

1 source
CivitaiPrimaryVersion 3099495
civitai.com/models/2754734

Verified Oct 18

Source excerpts

3 excerpts

Source context: 90 downloads · Type Workflows · Base model Qwen

Source context: 79 downloads · Type Workflows · Base model Qwen

Source context: 32 downloads · Type Workflows · Base model Qwen