Civitai #144478882: Retro Cel-Shaded Manga Boxcar - Indie Anime & Krea 2 Turbo

Full Resolution

Generation Specs RTX 4070

Sampler euler

Scheduler simple

Steps 10

CFG Scale 1

Seed 144478882

Target VRAM 12GB

VRAM 8GB - 12GB

GPU Compatibility

Optimal Performance

Runs entirely in VRAM with no memory swapping.

Krea 2 Turbo (Official Comfy-Org) Civitai #144478882 Min 8GB VRAM 1024×1024 RTX 4070 Verified

Retro Cel-Shaded Manga Boxcar - Indie Anime & Krea 2 Turbo

Vibrant retro cel-shaded manga art featuring a sun-drenched train boxcar overgrown with lush vines, expressive multi-character indie storytelling, and clean cel-shaded outlines synthesized on Krea 2 Turbo.

Blueprint Summary RTX 4070 Verified

Reproducible ComfyUI workflow for Civitai #144478882: Retro Cel-Shaded Manga Boxcar - Indie Anime & Krea 2 Turbo using Krea 2 Turbo (Official Comfy-Org) at 1024×1024 resolution. Requires minimum 8GB VRAM with sampler euler and scheduler simple (10 steps). Includes 1-click terminal model sync and canvas JSON graph.

Pipeline Flow 11 Nodes

Open in Resolver

01 Load Models

UNETLoader

5 loaders (DiT, CLIP, VAE)

02 Adapters

LoraLoader

2 adapters

03 Conditioning

CLIPTextEncode

2 prompt encodings

04 KSampler

KSampler

DiT latent denoising

05 Decode & Save

VAEDecode

Latent to pixel space

Execution DAG Topology Interactive Visualizer

Drag to pan · Scroll to zoom · Hover wires

Model & Asset Setup Setup Script

Run in your ComfyUI root:

curl -fsSL https://decomfy.com/api/scripts/civitai-retro-manga-boxcar-krea2.sh | bash

Positive Prompt

Retro cel-shaded manga art. A vintage blue boxcar overgrown with lush green vines against a sunlit yellow concrete wall. Standing on one leg on the boxcar roof at upper left, a young woman in denim overalls leans off the edge to reach for the string of a red balloon floating away; a young man in a white tee grabs the back of her strap to keep her from falling. At bottom left, another young man in a varsity jacket looks up, arms braced ready to catch her. At bottom right, a stylish young woman in an oversized Fishmans graphic t-shirt stands with hands on hips in the cool shadow of the boxcar sliding door, looking down at a young man, Jeff, stumbling backward below her. We see Jeff from behind in baggy jeans and sneakers as he stumbles backward, hands in the air. A cheerful young woman in a striped collared polo shirt peeks out an open train window at center frame, laughing joyfully at the scene. Beneath the window, detailed spray-paint graffiti of No-Face. At upper right, a retro folding lawn chair and striped parasol umbrella under clear blue summer sky, vibrant anime color palette, crisp ink outlines, cinematic composition.

Negative Prompt

unclothed, disrobed, revealing clothes, lowres, blurry, bad anatomy, deformed hands, extra fingers, missing fingers, low quality, artifacts, distorted proportions, text, watermark

Prompt Customization: This workflow is pre-wired and calibrated. Swap this prompt with your own character, subject, or style without breaking the pipeline.

LoRA Adapter Stack 2 Adapters

RTX 4070 Calibrated Weights

01 Mrpopo's Krea2 Anime/Art enhanced +1

02 Krea2 Refusal Reduction V2 (Full Rank) +1

Required Models 5 Models

Required Storage: 20.7 GB (5 models · DiT/Base: 12.6 GB · Text Encoder: 4.9 GB · LoRAs: 3.0 GB · VAE: 253 MB)

models/diffusion_models/ 12.56 GB HuggingFace

krea2_turbo_int8_convrot.safetensors Official Krea 2 Turbo INT8 ConvRot Diffusion Transformer core

Get

models/text_encoders/ 4.88 GB HuggingFace

qwen3vl_4b_fp8_scaled.safetensors Official Qwen3-VL 4B FP8 Scaled Text Encoder (CLIPLoader type: krea2)

Get

models/vae/ 253 MB HuggingFace

qwen_image_vae.safetensors Official 16-channel Video & Image VAE

Get

models/loras/ 545 MB HuggingFace (Mirror)

Mrpopo anime full potential.safetensors Weight: 1.0 (Mrpopo Krea 2 Anime/Art style enhancer with clean lineart and vibrant gouache tones)

Get

models/loras/ 2.48 GB HuggingFace (Mirror)

refusal_reduction_v2_full_rank.safetensors Weight: 1.0 (Full-rank Refusal Reduction LoRA for prompt fidelity and composition alignment)

Get

Field Notes RTX 4070 Benchmark

Benchmark: Rendered on NVIDIA GeForce RTX 4070 (12GB VRAM, CUDA 12.4)

1. ARCHITECTURAL REASONING & COMPOSITION GEOMETRY: Deconstructs complex multi-subject narrative staging within a retro cel-shaded anime aesthetic. Krea 2 Turbo naturally provides high-fidelity character rendering, but multi-character group composition often collapses into spatial incoherence. Chaining Mrpopo's Krea2 Anime/Art Enhanced LoRA anchors the vintage late-90s/early-2000s anime illustration style (crisp ink lineart, vibrant flat gouache shading, high-contrast sunlight), while the Refusal Reduction V2 full-rank adapter eliminates prompt clipping and allows rich environmental props (climbing ivy, vintage boxcar textures, No-Face graffiti mural) to materialize simultaneously. 2. CABLE ROUTING & SIGNAL TOPOLOGY: Dual sequential LoRA chaining topology over Krea 2 Turbo UNET (krea2_turbo_int8_convrot.safetensors) and Qwen3-VL 4B text encoder (qwen3vl_4b_fp8_scaled.safetensors). Both adapters modify diffusion model weights and CLIP embeddings before signal ingestion into conditioning nodes. A single deterministic Positive conditioning node is wired directly to KSampler, enforcing WYSIWYG canvas behavior without disconnected standby nodes. The latent output decodes cleanly through native qwen_image_vae.safetensors. 3. HARDWARE BENCHMARKS & CONSUMER GPU FOOTPRINT: Executed on host NVIDIA GeForce RTX 4070 (12GB VRAM, CUDA 12.4). Peak VRAM usage reaches ~10.2 GB during the 10-step Euler generation, well within the 12GB threshold. Generation finishes in ~60-70 seconds without paging into host system RAM.

Frequently Asked Questions FAQ

What GPU and VRAM are required to run Civitai #144478882: Retro Cel-Shaded Manga Boxcar - Indie Anime & Krea 2 Turbo?

This workflow requires a minimum of 8GB VRAM (recommended 12GB VRAM). Tested and verified on NVIDIA GeForce RTX 4070 (12GB VRAM) at 1024x1024 resolution.

How do I resolve missing custom nodes for this workflow?

You can drop the workflow JSON into our client-side Missing Node Auto-Resolver at https://decomfy.com/resolve/ to detect missing nodes and generate install commands, or run the 1-click terminal setup script provided below.

Which VAE and conditioning text encoders are required for Krea 2 DiT?

Krea 2 DiT single-stream models strictly require the 16-channel Wan2.1 VAE and Qwen3-VL text encoder. Using standard Flux or SDXL VAE causes inverted checkerboard artifacts. Also pair with RescaleCFG for high-contrast portrait clarity.