Skip to content
WorkflowListedExternal
Download

Blink's universal workflow | CivitAI metadata, upscaling, FaceDetailer (+indexing elements), inpainting, controlnet, IPAdapter (regional conditioning) and token counter (SDXL/PONYV6/ILLUSTRIOUS/NOOBAI-XL/ANIMAGINEXL4/SD1.5) ComfyUI beginner friendly all in one — v5

Cleaned up, added i2i.

Sourcecivitai.com/models/1621455
VersionRelease 3394247 — blinksAIOSDXLWorkflow_v6.jsonEstimated VRAMNot yet characterizedUnratedgrade

Runtime profile

Access
External workflow
Status
Listed
Version
blinksAIOSDXLWorkflow_v6.json
Source
Civitai
Updated
Nov 3, 2026

What you need

Hardware and download-size estimates are not available yet.

Requirements

Models

9

  • 4x-UltraSharp.safetensors

    Repository matchedOptional

    Upscaler · Hugging Face · Kim2091/UltraSharp

  • CLIP Vision (ViT-H).safetensors

    Repository matchedOptional

    Clip Vision · Hugging Face · Kim2091/UltraSharp

  • controlnet-union-sdxl-1.0-promax.safetensors

    Repository matchedOptional

    ControlNet · Hugging Face · brad-twinkl/controlnet-union-sdxl-1.0-promax

  • ema_vae_fp16.safetensors

    Not resolvedOptional

    VAE · Unknown

  • face_yolov8m.pt

    Not resolvedOptional

    Object Detector · Unknown

  • ip-adapter-plus_sdxl_vit-h.safetensors

    FoundOptional

    LoRA · 808 MB · SAFETENSORS · FP32 (unverified) · Hugging Face · h94/IP-Adapter

  • sam_vit_b_01ec64.pth

    Repository matchedOptional

    Segmentation · Unknown

  • seedvr2_ema_7b-Q4_K_M.gguf

    Not resolvedOptional

    Checkpoint · Unknown

  • Soft Surrender vPred v9.safetensors

    Not resolved

    Checkpoint · Unknown

Node packs

25

  • Any Switch (rgthree)

    Not resolvedOptional

    Node pack · Unknown

  • Bookmark (rgthree)

    Not resolved

    Node pack · Unknown

  • ComfyUI_Comfyroll_CustomNodes

    Found

    Node pack · GitHub · Suzie1/ComfyUI_Comfyroll_CustomNodes

  • comfyui_controlnet_aux

    PossibleOptional

    Node pack · Registry

  • comfyui_ipadapter_plus

    PossibleOptional

    Node pack · Registry

  • comfyui-image-saver

    Found

    Node pack · GitHub · alexopus/ComfyUI-Image-Saver

  • comfyui-impact-pack

    Found

    Node pack · Registry

  • comfyui-inpaint-cropandstitch

    FoundOptional

    Node pack · Registry

  • DetailDaemonSamplerNode

    FoundOptional

    Node pack · Unknown

  • easy ifElse

    FoundOptional

    Node pack · Unknown

  • easy imageScaleDownBy

    FoundOptional

    Node pack · Unknown

  • Fast Groups Bypasser (rgthree)

    Not resolved

    Node pack · Unknown

  • ImageResizeKJv2

    Possible

    Node pack · Unknown

  • ImpactSwitch

    FoundOptional

    Node pack · Unknown

  • Label (rgthree)

    Not resolved

    Node pack · Unknown

  • Ryuu_TokenCountTextBox

    Found

    Node pack · Unknown

  • SAMLoader

    FoundOptional

    Node pack · Unknown

  • Sampler Selector (Image Saver)

    Found

    Node pack · Unknown

  • Scheduler Selector (Image Saver)

    Found

    Node pack · Unknown

  • SeedVR2LoadDiTModel

    PossibleOptional

    Node pack · Unknown

  • SeedVR2LoadVAEModel

    PossibleOptional

    Node pack · Unknown

  • SeedVR2VideoUpscaler

    PossibleOptional

    Node pack · Unknown

  • UltimateSDUpscaleCustomSample

    PossibleOptional

    Node pack · Unknown

  • UltralyticsDetectorProvider

    PossibleOptional

    Node pack · Unknown

  • VRAM_Debug

    FoundOptional

    Node pack · Unknown

AI-generated commentary

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

Purpose

  • Use this ComfyUI workflow to create or edit images from text prompts or a loaded image and mask, with options for upscaling, FaceDetailer, inpainting, ControlNet, regional IPAdapter conditioning, and token counting; image results can be previewed or saved.

Inputs and outputs

  • For inpainting, the notes describe copying an image into Load Image, painting a mask, and saving it; the graph also contains PreviewImage and Image Saver nodes.

Setup

  • For LoRAs, select them in the wildcard processor and copy them into the text field; enable FaceDetailer, upscaling, or ControlNet with Enable Group in Fast Groups Bypasser (rgthree), at the top left.

  • Use shortcuts 1–4 to navigate the workflow; 3D posing requires a browser with hardware acceleration.

Dependencies

  • The workflow lists these required node names: Bookmark (rgthree), ComfyUI_Comfyroll_CustomNodes, comfyui-image-saver, comfyui-impact-pack, Fast Groups Bypasser (rgthree), ImageResizeKJv2, Label (rgthree), Ryuu_TokenCountTextBox, Sampler Selector (Image Saver), and Schedulerl

  • Optional node names include Any Switch (rgthree), comfyui_controlnet_aux, comfyui_ipadapter_plus, comfyui-inpaint-cropandstitch, DetailDaemonSamplerNode, easy ifElse, easy imageScaleDownBy, ImpactSwitch, SAMLoader, SeedVR2LoadDiTModel, SeedVR2LoadVAEModel, SeedVR2VideoUpscaler,、

  • Soft Surrender vPred v9.safetensors is listed as the required model file; optional model files are 4x-UltraSharp.safetensors, CLIP Vision (ViT-H).safetensors, controlnet-union-sdxl-1.0-promax.safetensors, ema_vae_fp16.safetensors, face_yolov8m.pt, ip-adapter-plus_sdxl_vit-h.safet

Configuration

  • The included note suggests Rescale CFG 0.3 for anime, 0.7 for CFG 7–15, or 0 as neutral.

    Suggestion · not verified

Cautions

  • The author keeps FreeU_V2 bypassed because they say it does not work well with anime.

  • A license is not specified in the listing; check the publisher's page for usage terms before using the workflow.

    Suggestion · not verified

Does this need to be edited?

Sign in to send an edit request.

Sources

1 source
CivitaiPrimaryVersion 3394247
civitai.com/models/1621455

Verified Nov 3

Workflow instructions

Author notes from blinksAIOSDXLWorkflow_v6.json · selected release

SEGS Filter (ordered) Here you choose how many heads/hands/persons.. you want to change.

Author notes from blinksAIOSDXLWorkflow_v6.json · selected release

Zero Star CFG +Improves how well the generated image matches your text prompt. Rescale CFG 0.3 for anime 0.7 should work well for CFG 7-15 0 is neutral. +By balancing the CFG scale, you can reduce artifacts and achieve more natural-looking images. FreeU v2 +Improves details and reduces artifacts in generated images. -Optimal results require adjusting multiple parameters b1: Controls the balance between the backbone and skip connections in the U-Net architecture. b2: Adjusts the strength of the backbone's influence. s1: Modulates the skip connection's contribution to the final image. s2: Fine-tunes the overall blending of features. FreeU_V2 values SD1.5: b1: 1.5, b2: 1.6, s1: 0.9, s2: 0.2 SDXL: b1: 1.3, b2: 1.4, s1: 0.9, s2: 0.2 https://github.com/ChenyangSi/FreeU 1 is neutral. Self-Attention Guidance (SAG) +Focuses on important areas of the image, leading to finer details. -Adds extra processing steps, which might slow down generation. -Effectiveness can vary depending on the underlying model architecture.

Author notes from blinksAIOSDXLWorkflow_v6.json · selected release

>RMB on text "image_a" in the Image Comparer node >Copy Image >Paste (ctrl + V) generated image in the Load Image node (above this note) If it creates a new node delete it and set name to pasted/image in existing one >RMB, open in mask editor >Paint a mask, make sure to set max opacity >Save You can just make outlines, mask_fill_holes is on. Inpainting small area, AI needs more context? Increase context_from_mask_extend_factor.

Author notes from blinksAIOSDXLWorkflow_v6.json · selected release

number of characters (1girl/1boy), character name, artist prompt, scene/environment/camera angle, action, expression, items, quality prompts (masterpiece, best quality) huke, ilya kuvshinov, saitom, yoshida akihiko \(style\), wlop, ross tran

Download files

Model artifacts

2 artifacts

blinksAIOSDXLWorkflow_v6.json

json · 167 KB · SHA-256 47d76fb317da…fe11 · Civitai

Download file

blinksAIOSDXLWorkflow_v6.json

json · 167 KB · SHA-256 47d76fb317da…fe11

Download file

Source excerpts

1 excerpt
NoteNov 3civitai.com/models/1621455

Source context: 2092 downloads · Type Workflows · Base model NoobAI