Anima Reverse Prompt Simplified Workflow (Image → Tags → Anima Format) (6G VRAM can run it!)

Anima Reverse Prompt Simplified Workflow (6G VRAM can run it!)

Drop an image → auto reverse prompt → auto classify tags → output already in Anima format. No manual tagging.

🎮 Run it online (no install)

Global: https://www.runninghub.ai/user-center/1898017693566418945/webapp?inviteCode=rh-v1315

Click the link above to try it without downloading any model. If the results are good, deploy it to your local environment.

Register and bind my invitation code rh-v1315 to get an extra 1000 RH coins. You also get 100 points by logging in daily — usable for both image and video generation.

Video walkthrough:

📖 Full documentation — read this first

Node-by-node explanation, install steps, Python dependencies, llama.cpp setup, model placement, and how to fix red nodes:
https://icn01wg02kz3.feishu.cn/wiki/GrCtwn5tSignWokefrycOxe8nNf?from=from_copylink

📦 Required plugins (8)

1. ComfyUI-Danbooru-Anima-Prompt — the plugin this workflow is built for
https://github.com/Zove-try/ComfyUI-Danbooru-Anima-Prompt

2. ComfyUI-TaggerPlus — fast WD14 / PixAI tagger
https://github.com/Zove-try/ComfyUI-TaggerPlus

3. comfyUI-llama-TE — local Qwen3.5-VL reverse prompt
https://github.com/tl2012tl/comfyUI-llama-TE

4. Comfyui_Comfly
https://github.com/ainewsto/Comfyui_Comfly

5. rgthree-comfy
https://github.com/rgthree/rgthree-comfy

6. ComfyUI-Easy-Use
https://github.com/yolain/ComfyUI-Easy-Use

7. ComfyUI-KJNodes
https://github.com/kijai/ComfyUI-KJNodes

8. ComfyUI-Lora-Manager
https://github.com/willmiao/ComfyUI-Lora-Manager

🧩 Models used (4)

1. Anima Base v1.0 — the main model (2B, anime / illustration)
https://huggingface.co/circlestone-labs/Anima
All three files below live in the repo's split_files/ folder.
anima-base-v1.0.safetensors  →  ComfyUI/models/diffusion_models/
qwen_3_06b_base.safetensors  →  ComfyUI/models/text_encoders/
qwen_image_vae.safetensors  →  ComfyUI/models/vae/

2. Anima Turbo LoRA v0.1 — speed LoRA, cuts sampling to 4–8 steps
https://civitai.com/models/2560840/anima-turbo-lora
anima-turbo-lora-v0.1.safetensors  →  ComfyUI/models/loras/

3. Qwen3.5-9B-Uncensored-HauhauCS-Aggressive (GGUF) — local reverse prompting (Qwen TE line)
https://huggingface.co/botp/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive
Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-Q6_K.gguf
mmproj-Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-BF16.gguf
 →  both go in ComfyUI/models/LLM/

4. PixAI Tagger v1.0 — tagging model used by ComfyUI-TaggerPlus
https://huggingface.co/pixai-labs/pixai-tagger-v1.0
 →  ComfyUI/models/pixai_tagger/

All of the above are already inside the Quark pack — if you grab the pack you don't need any of these links.

Anima weights are released under CircleStone Labs' non-commercial license — check the repo before commercial use.
📌 Two versions included

V1.0 (Gelbooru) — uses the GelbooruBrowser node
V1.0 (Danbooru) — uses the DanbooruBrowser node

Everything else is identical. Pick whichever tag source you use more.

Sign up via this link claim 1000 RH Coins and generate tons of images/videos for FREE!
My AI models/apps are on RH now! Try them out & support my work — every click counts. Thank you!
Thank you all for your support. Your support is the motivation for me to update more interesting and useful workflows. Thank you so much!!!