ComfyUI Workflows
Browse real AI workflows from Flux, SDXL, Pony, ControlNet, video generation and commercial AI creators.
Illustrious/ Pony/ SDXL SFW & NSFW Pro Grade Workflow (Low and high VRAM)
Note: I spend a lot of time on these to make sure they are top notch. If you have ANY issues, I'm very responsive. hit me up and I'll resolve them for you.This is an advanced workflow. It is node heavy and designed for multiple options and to return professional high quality results. If you are looking for simpler workflows, they are here:An essentials one can be found here: https://civitai.com/models/2349309/illustriousponysdxl-essentialsA simple one can be found here:https://civitai.red/models/2600919/lonecats-simple-zitilluanimaflux-workflows?modelVersionId=2921603A complete user and troubleshooting guide inclusing a list of common error messages can be found here https://civitai.com/articles/24678/how-to-use-my-workflows-a-comprehensive-breakdownInstagram: https://www.instagram.com/synth.studio.models/Buy me a☕ https://ko-fi.com/lonecatoneThis represents hundreds of hours of work. If you enjoy it, please 👍like, 💬 comment , and feel free to ⚡tip 😉也有中文说明VER 10.1:Got tired of having to always mess with resolution, so I gave you the option to make it automaticMath node replacements to deal with the Linux errorsAdded psst... fast bypassers (you'll thank me).💡use the upgraded face yolo I added. much more accurate.Improved notes throughoutVer 8.0/ 9.0:Major overhaul to bring it up to speed with my other workflowsAdded second checkpoint metadataadded a prompt to hi-res fixadded Inpainting suite from the pro grade workflow by request.V7.0Apparantly , I forgot to hook up Face Swap in my last model and nobody told me until today, so a free major upgrade it is!Installed a sentence remover in the prompt assist area to help eliminate CivitAI 🚩Cleaned up some bugsResinstalled Qwen VL Prompt EnhancerUpgraded Face Swapalso upgraded notes on how to remove it.Minor changes to detailersAdded an inline crop tool in the post production suite.Ver 6.0:Major overhaul to bring it up to speed with my other workflowsAdded a refiner checkpoint that can easily be switched on or offAdded FaceID as well as a style and composition IP Adapter for character consistencyReowked the Detailers ro improve thembetter version of Hi Rez fixReworked the Save with Metadata to align with CivitAi and fix the Comfy update bug 🙄This is an all in one workflow:ControlnetI2I with prompt generationDual Model Loader: Better controlSage Attention: Speeds up productionUltimate upscaler: For prescale and hi-res fixDetailers: Face, eyes, hand, expressionNSFW Detailers: Male & Female anatomySeed VR2 Upscaler: Get up to 4k pics, even on low VRAMPost production suite: My pride and joy. Does everything.
txt2img Prompt Assist Randomizer
Just like when you're looking for a recipe and you have to read through a 5 paragraph story about someone tasting some dish in another country and and how it reminded them of home, im going to put you guys through the same BS so you know why i made this nightmare prompt assistant. the last week ive been working on a new lora for illustrious and when i was going through it i found out if you shove abunch of wildcard style prompting into a text encode it seemed to work, i got multiple poses, camera angles, and as well as the thing i was training the lora on (vist the profile later or follow for updates). Well i thought why dont i make my own little prompt assistant tool that can help other people train loras as well? originally i was going to gate keep this to myself but after a friend asked if they could use my workflow as a reference for another workflow they were working on i thought well damn if he wants to use this then maybe other people would too. so here you guys are! the fruits of my madness using my original EZtxt2img workflow with a better upscaler than prior workflows. if you made it this far and those who havent checked out my other workflows here is the sales pitch for this one.ARE YOU TIRED OF WRITING THE SAME PROMPTS OVER AND OVER?INTRODUCING THE WILDCARD SELECTOR WORKFLOW — THE ULTIMATE PROMPT RANDOMIZER FOR COMFYUI!Featuring over 60+ INDIVIDUALLY TOGGLEABLE NODES across 17 CATEGORY GROUPS, this workflow puts TOTAL CONTROL at your fingertips!BUT WAIT, THERE'S MORE!EYES — 5 dedicated randomizers for color, shape, pupil shape, pupil color, and sclera! Want heterochromia with slit pupils and black sclera? FLIP A SWITCH!FACE — 5 randomizers covering face shape, details, expression, gaze direction, and makeup! NEVER GET SAMEFACE AGAIN!HAIR — 5 randomizers for color, length, style, bangs, and accessories! OVER 500 POSSIBLE COMBINATIONS!BODY — Type, skin, and detail controls! FROM PETITE TO MUSCULAR, WE GOT IT ALL!13 CLOTHING STYLES — Fantasy! Sci-Fi! Techwear! Gothic! Victorian! Cosplay! Office! Lingerie! Girlfriend! Vintage Housewife! AND MORE! Each one INDIVIDUALLY TOGGLEABLE!11 KEMONOMIMI SPECIES — Cat! Dog! Wolf! Fox! Rabbit! Deer! Horse! Dragon! MIX AND MATCH for chimera combos!7 OTHER SPECIES — Elf! Human! Succubus! Fairy! Ghost! Slime Girl! Goblin! Because why not!DUAL ENVIRONMENT MODES — Simple backgrounds for TRAINING DATASETS or detailed backgrounds for SHOWCASE GENERATION! Switch between them INSTANTLY!BODY VISIBILITY CONTROLS — Coverage level, anatomy emphasis, and exposure area toggles for building CLEAN TRAINING DATASETS! Get those attachment points and transition zones your LoRA ACTUALLY NEEDS!POSE & CAMERA — Pose, camera angle, and composition randomizers! Every gen a new angle!5 POSITIVE PROMPT BOOSTERS — Quality, color palette, material finish, mood, and art style!5 NEGATIVE PROMPT TEMPLATES — Quality cleanup, style control, anti-chimera, kemonomimi cleanup, and body cleanup! ALL TOGGLEABLE!POWERED BY IMPACT PACK'S WILDCARDPROCESSOR — ACTUAL randomization, not just vibes!COMPATIBLE WITH SDXL / ILLUSTRIOUS CHECKPOINTS!PERFECT FOR:LoRA dataset generationStyle explorationBatch variety generationPeople who are too lazy to write promptsOVER 10,000+ POSSIBLE PROMPT COMBINATIONS FROM A SINGLE WORKFLOW!ORDER NOW AND WE'LL THROW IN A FREE SET OF NEGATIVE PROMPTS ABSOLUTELY FREE!results may vary, not responsible for cursed generations, extra fingers sold separatelymay require additional LoRAs for optimal resultscreator is not liable for any unholy chimera abominations generated with the kemonomimi mixerthe "girlfriend cozy" preset will not get you an actual girlfriendside effects may include: prompt addiction, uncontrollable batch generation, hoarding 500GB of anime girls on your hard drive, and explaining to your power company why your electricity bill tripledsuccubus species toggle has been known to cause users to forget they were building a training datasetvintage housewife preset does not include cooking skillsbody visibility section is for SCIENCE and TRAINING PURPOSES ONLY wink winknot responsible for any GPUs harmed in the making of this workflowif your generations look cursed for more than 4 hours please consult a different checkpointcompatible with most SDXL/Illustrious checkpoints, your mom's laptop is NOT a supported platformno waifus were harmed in the development of this workflowby downloading this workflow you agree that cat ears improve everything
Eazy Regional Prompter - Illustrious Workflow for ComfyUI
Easy Regional v0.6 - (Please read the Prompt Guide & installation requirements)This ComfyUI workflow is designed for SDXL/Illustrious models, with built-in support for multi-character image generation. It features an all-in-one solution, with txt2img > hires fix > face detailer. The primary goal of this workflow is to provide an intuitive, easy-to-use UI that mirrors A1111 / Invoke, while remaining flexible for advanced use cases. The final image will include embedded metadata to make uploading to Civitai easier.Regional LoRA Support!Regional LoRAs are now supported. To add a LoRA to a region, type <lora:LORA_NAME:STRENGTH> [TRIGGER_WORD] into the region you wish the LoRA to apply to eg: (works best with characters)<lora:lauma_genshin_impact_ilxl_goofy:1.0> lauma \(genshin impact\)Prompt GuideCommon Prompt: Tags shared between regions such as, quality and background tags. Use this to define the amount of people and their gender eg. 2girls or 1boy, 1girl etc.masterpiece, best quality, absurdres, outdoors, park, bench, sunny, sunrays, 2girls, front view, straight onRegion 1: First character and/or specific details to this region (start this prompt with the gender of the character male or female NOT 1girl or 1boy)female, hatsune miku, aqua eyes, long hair, twin tails, hair orniment, white shirt, collard shirt, bare shoulders, black sleeves, pleated miniskirt, black thighhighs, sittingRegion 2: Second character and/or specific details to this region (start this prompt with the gender of the character)female, frieren, green eyes, pointy ears, twin tails, red earrings, capelet, striped top, white skirt, cuddlingNegative prompt: Standard negative prompt tags, nothing fancy going on herelowres, worst quality, bad quality, watermark, signatureSuggested Resolutions:1024 x 1024896 x 1152832 x 12161152 x 8961216 x 832Installation RequirementsIf the input fields on the nodes are in the incorrect place after installing the custom nodes, please redownload the workflow and the connections should be correct.This workflow requires the following model types:Base model: ComfyUI/models/checkpointsVAE eg. sdxl-vae-fp16-fix: ComfyUI/models/vaeUpscale Model eg. remacri: ComfyUI/models/upscale_modelsThis workflow also requires these node packs in order to function, these can be easily installed via the built-in Custom Nodes Manager:
Illustrious/ Pony/ SDXL Pro Grade Workflow (Low or High VRAM)
Note: I spend a lot of time on these to make sure they are top notch. If you have ANY issues, I'm very responsive. hit me up and I'll resolve them for you.This is an advanced workflow designed for people familar with Comfyui and who want pro grade results. I encourage you to learn it, but if it is overwhelming, the simple version can be found HEREA complete user and troubleshooting guide inclusing a list of common error messages can be found here https://civitai.com/articles/24678/how-to-use-my-workflows-a-comprehensive-breakdown也有中文说明Instagram: https://www.instagram.com/synth.studio.models/Buy me a☕ https://ko-fi.com/lonecatoneThis represents hundreds of hours of work. If you enjoy it, please 👍like, 💬 comment , and feel free to ⚡tip 😉V15.0Because it is not in my nature to leave things alone....Gutted the stock face detailer nodes and replaced them with a SEGS suite for infinately better resultsNote: Yes, I will be going through all my workflows and replacing them, including teh NSFW onesControlnet Update.Added local bypass nodes so ou don't have to zoom and scrollAdded a quality modifier prompt area that follows after Qwen so that they stay preserved.Major update to notes, especially for detailersV12.1Added an Inpaint areaModified controlnet so it doesn't fault out if a mask isn't loaded without resetting teh switchRemoved the sentence remover (it sucked)added "banned tag remover" at the metadata saver so you can still process normallyV11.1.2Updated 📝Prompt AssistUpdated ⚙️ main areaAdded back Qwen VL prompt enhancer (now that it's fixed)Updated 🔎 detailersUpdated 💾 Save area so that you can embed workflows as well as add a simple date to the folders so you aren't endlessly looking for them.Added an inline ✂️ crop tool in the 🎨Post Production SuiteVer 10Improved regional prompt suitesAdded IP Adapters for Face, style, and composition for Facial and pose consistencyImproved detailer suitereworked settings for easier useAdded CivitAI sentence ✂️Ver 9.0Qwen nodes keep getting mbroken due to updates. I removed them in V9Removed outdated sectionsAdded a 2nd KSampler & model option for refinementV 8.1 is included in the folderVER 8Created a really cool regional prompter.Works with controlnetUpgraded detailers to accomodate for it so you can detail each face seperatelyThis is an all in one workflow:Controlnet: background remove, canny, openpose, depth, canny/depthI2I prompt assistQwenV2 Prompt enhancer.Florence and Qwen prompt assistStyle selectorDual Model Loader: Better controlSage Attention: Speeds up productionUltimate Upscaler for prescaling and hi res fixDetailing Suite for face, hands, eyes, and expressionsmulti face detailerSeed VR2 Upscaler to get 4k on low VRAMMassive Post Production SuiteSmart Noise scrubberFull tutorial available.也有中文说明
Text2Img Workflow with Detailer Arrays
Hello Everyone!Have you ever created an image you like and would have kept, but and a face, or hand just came out wrong? You need a workflow that can take control address that.Ever make an image that had multiple characters, but when you tried to assign them the correct eye color or expression, the checkpoint got confused, and you had to play generation roulette? You need a workflow that lets you decide whose face has just came out like what.That is what this workflow, used properly, will allow you to do. This workflow is my daily driver T2I workflow for checkpoints based in SDXL, Illustrious and Pony. It is based upon my Img2Img SDXL Workflow + META, which I was only able to create thanks to bits taken from other workflows, particularly those of Legendaer and Yukichan_, and advice from Lonecatone23. Over several revisions, I have made a number of refinements to the workflow while at the same time striving to keep it simple so that even a beginner can take hold of this workflow and start generating.Main Features:✦ Text to Img Generation - That's what we're here for, text to image.Include Metadata / Embed Workflow Option. Stable 1 megapixel generation, or dare to increase it with Size Boost.✦ Metadata saving included — CivitAI ready. You can also embed the entire workflow into your images, by default this is turned off.✦ Upscaler with Color Matching - this is the feature that makes images go from good to great. Or use HiresFix for a quick increase in resolution✦ Quality Tags Presets Switch - instead of typing Masterpiece, best quality, etc. etc.✦ Two Segs Detailer Arrays - two SEGS pipelines (by default, hands and faces) that you can individually prompt or apply one prompt to all. Sophisticated tool created to handle individual details - your backstop for when the image has to be perfect.✦ Parallel development with I2I workflow - if you learn this one my Img2Img works essentially the same way.✦ CFG Controls - need more control, more CFG? It's handled automatically, crank it up to 10.✦ LoRA Manager — Search your LoRA database with style, I love this node. Dump it if you prefer something else.✦ Simple and reliable — Enter your prompt, hit run, let the segs previews populate, freeze the gen seed, and fill in prompts and run it again. That's it.New with 5.0 SESkipped 4.0, just becauseSeed JumpStyle Preset switchingAdded a toggle and Subgraph for exceeding native SDXL resolution on initial genFixed Typos for name of Face Detailer LoRA blocker nodesDoes not include Quality Tags Fast Switching, uses native concatenate nodes5.0.1 Changes:Changed Positive Text Box to Mira, which was already being used in this workflowChanged Wildcard ProcessorSeed controls Wildcard selection (Freezing seed freezes your wildcards too now)Added Final Prompt PreviewAdded a toggle and Subgraph for exceeding native SDXL resolution on initial genSlight rearrangement of Control Group, hopefully optimalNew with 3.0:With 3.0, the layout has changed and metadata flows have changed. The checkpoint loader is now the checkpoint selector, which should provide some convenience and added stability. The LoRA blocker no longer requires fiddling with the spaghetti string to toggle it, there is now a toggle switch. Sliders have been implemented throughout the Input center, including controls for the Upscaler.New with 2.2:Added toggles for including the workflow in the image's metadata.Added a means by which you can bypass the upscaler and still send the image (directly from phase 1 gen/hires fix) to the detailer, rather than being forced to use upscaler in order to access the detailer.Color Matching directly following Upscaler - I have found that some models have noticeable drift in color tone and this corrects that, works beautifully.New with 2.1:Detailer moved to the back of the order, after upscale, to ensure that full control over expressions remains with you, the artist. I'm very happy with this workflow in its current form, although there might be new versions or changes down the line.New with 2.0:Raise CFG without crushing the imageUpscaler with a third save nodeSomehow resists LoRA bleed, whether it's the scheduler / sampler combo or what I'd love to know. It's not perfect though!Preview KSampler - a game changer for skipping unsatisfactory seeds. Use it!You will need to download an upscaler model if you don't have one already.How this Workflow Came to BeI was using one of these massive workflows with all the bells and whistles and I ran into a snag. A LoRA was confounding the hand detailer and the hands were all coming out like macaroni alfredo. So, I turned off that LoRA. I was also running into a common problem: when two characters are in the frame, they end up with the same eye or hair color. The tags were blending across character BREAKs which is ultra common, and detailers that handle all bboxs in order cannot be instructed on each bbox individually. That's something that my I2I workflow was designed to do, and the mega huge T2I workflow with all its heft didn't have detailer separation as a feature.I don't think I could add to such a complex workflow without tripping up and it end up taking 2 hours. I thought if I'm going to modify a workflow it might as well be my own. It would keep things simple with only nodes I know I have, I'd finish in under 30 minutes, and then I would have complete control over faces.Saving the metadata is important too, however somehow in converting things and removing things that get concatenated (such as the natural language prompt), the metadata portion got a little scrambulated, and don't worry I fixed it -- but 30 minutes turned into a several hour affair. Under nine hours though, I'm pretty sure. The latter part was mostly testing and tweaking. I ended up making some changes to the detailer modules so that they don't spam the image feed, and upgraded that module to V1.1.During testing, I started with one figure, then two, and I soon discovered that this workflow is absolutely a beast at ensuring your faces all look great. Whether you have one face in the image or five, the detailer can make each one pop with high aesthetic quality and unique expressions, and the SEGS previewer keeps you in control.The hand detailer gives you per-segs control over hands although sometimes bad hands are going to happen, particularly when there are many in one image. Fingers in AI is always a little trickier, but the workflow allows you to add a hand detailer LoRA directly to the detailer, upping your chances of success dramatically. Do it. At the same time, that LoRA will be blocking other LoRAs that could create havoc on delicate fingers. I haven't seen this approach elsewhere, curious whether it holds up in your experience. Please let me know in the comments what you think about my LoRA blocker, if you have any aspects you think ought to be changed, bug fixes, etc. This is just version 1, it's bound to evolve.RequirementsYou will need a face bbox detector - like this oneYou will needs a hands bbox detector - like this oneYou might decide to add or replace a bbox detector for face or hands with one for eyes, or add another bbox array. For eyes, try this oneIn ComfyUI you may need to turn some safety features off to use pickletensors. Consult AI if you run into problems.Custom Nodes Used:rgthree-comfyComfyUI Impact PackComfyUI Impact SubpackComfyUI-Easy-UseComfyUI-Lora-ManagerComfyUI Image SaverComfyUI_essentialsComfyUI-KJNodescg use everywhere Anything EverywhereDynamic ThresholdingComfyui_MiraSkimmed_cfgefficiency-nodes-comfyuiComfyroll StudioComfyMath
Eazy Illustrious - Workflow for ComfyUI
Easy Illustrious Workflow v0.4 - (See installation requirements)This ComfyUI workflow is designed for SDXL/Illustrious models. It features an all-in-one solution, with txt2img > hires fix > face detailer. The primary goal of this workflow is to provide an intuitive, easy-to-use UI that mirrors A1111 / Invoke, while remaining flexible for advanced use cases. The final image will include embedded metadata to make uploading to Civitai easier.Positive prompts now support wildcards and dynamic tags!Wildcards: Reference files from your ComfyUI custom nodes directory by enclosing the filename in double underscores (e.g., __filename__).Dynamic Tags: Create instant random variations directly in your prompt using curly brackets separated by vertical bars, like {blue|red} ball or {sitting|standing}.Wildcard preview: Use the 'Preview Wildcard' node to view the final prompt generated by your wildcards and dynamic tags.Installation RequirementsThis workflow requires the following model types:Base model: ComfyUI/models/checkpointsVAE eg. sdxl-vae-fp16-fix: ComfyUI/models/vaeUpscale Model eg. remacri: ComfyUI/models/upscale_modelsThis workflow also requires these node packs in order to function, these can be easily installed via the built-in Custom Nodes Manager:
Silly All-in-one, Multiple Characters With Regional Loras
This ComfyUI workflow was mainly created to do one thing in SDXL: "Multiple characters Loras" that embeds every process of how the image is generated.It's like semi-inpaint that starts from the beginning of the generative process, and is compatible with "Tile Upscale (Ultimate SD Upscale)".It also works well without character lora.It works well with AMD GPU, should work with an Nvidia GPU too, if the custom nodes work as intended.📌FeaturingIllustrious 2.0 Focused Workflow (Supports SDXL models)Regional Prompts With Regional Lora (Up to 3 characters)All-in-one Workflow (KSampler/Ultimate SD Upscaler/Detailer/Color Correction/Upscale/Inpaint)Optional ControlNets (OpenPose, DepthAnything, Canny, SoftEdge, Tile)Separated Base/Detailer Prompts (less upscaler/detailer hallucinate)Flexible Workflow Starts (Starts your generation from any part of the process)Semi-Auto Mask (starts workflow right away without manual masking involvement)Suitable for users who can navigate the workflow spaghetti.However, I've arranged the node groups in order from left to right and up and down. Categorized separately by Header and Textboxes. It should make it easier to follow through and input text/number accordingly. With Hotkeys to quickly navigate the workflowThe Guide, how to use, and that sort of stuff are inside the workflow. Ask me if there are any confusing words in the guide that may need improvement in the comments.Please avoid flooding me with personal chat as best as you can. I encourage posting your question in the comments.⚠️ForewarningBackup your Local ComfyUI Environment.When installing/updating new custom nodes, things can break from forcing to install certain version of site-packages, which may be incompatible with your settings.Not suitable for impatient users. The workflow is very delayed.Depending on how many group toggles are enabled.And if using "Regional Prompt/Lora", it will generate very slowly.2-4 times slower (depending on how many characters/character loras enabled).There's a basic auto masking, but complicated poses still require manual masking.Similar to "Inpainting", it requires tracing the character from the image to give proper facial expression/pose and prevent character features from bleeding.💾Download Section🎛️ControlNet (models > controlnet)OpenPose XL v2Depth Anything - Rename file to "t2i-adapter-depth-midas-sdxl-1.0".Canny - Rename file to "t2i-adapter-canny-sdxl-1.0".Soft EdgeTTPLanet Tile🔍Detailer (models > ultralytics)📦BBOX (ultralytics > bbox)Face yolov8n v2Eyeful v2 Paired or Eyeful v2 Individual (Can use either Eye Detector models)Person Yolo8m SEGS📐SEGMENT (ultralytics > segm)PitHandDetailer-v2-Test-v9c2d Mouth🧠Sams Detector (models > sams)sam_vit_b_01ec64💎Upscale ModelAnimeSharpV4_Fast_RCAN_PU⚠️Known Issue[If using Load Image node] You need to set your image "Aspect Ratio" to match your "Load Image". Still trying to fix the dependency cycles error when implementing "Auto Get Image Size".VERY LAGGY/DELAYED when generating due to the sheer amount of nodes and checks on the frontend. Can be remedied by keeping the browser size small.At the last part of every prompt. You have to manually add a comma "," at the end.Overburn/Overbaked Image could still happen if using 3 characters Loras, the countermeasure is to lower CFG by 1-3 for KSampler or use Sampler: "Euler_a_cfg_pp" with CFG: 1.3-1.5.Somehow Sampler: "Euler_a_cfg_pp" in Detailer causes artifacting in the corner, hence I use the standard Sampler/Scheduler combo for this part[Detailer "Tile SEGS" and Detailer Fix's "Character Bound" Mode] It is somehow forcing the generation outside of the detected area. It still works, but takes unnecessary long generative time.CreditsNerdy Rodent (Video Link)His video sparks my interest in creating this madness AIO of workflow using his method.Not so simple (or is it?) Illustrious Workflow by @Gladastook some good sampler/scheduler combo, nodes, and settings from thereRegional Conditioning with AttentionCouple(LoRA supported) by @Nyraen07Took some inspiration from there too. Also works similarly to mine but with different custom nodes (uses "A8R8 Attention Coupling" + built-in "Set Cond Props" nodes, my workflow only uses ComfyUI built-in "Set Cond Props + Set Default Cond Props")END NOTEI took an interest in generating stuff that I'm... interested in, along with gaining some useful insight, and sharing images/videos. This workflow is part of my learning and experimenting node and workflows.If you find this workflow on another site besides Civitai, the uploader is responsible for providing the support.I have no way of knowing what gets changed by the uploader.May you find my personal workflow useful and suits your needs.
PERFECT_image-workflow_TripTrap-V777-working
The amount of things this workflow has packed-in would quite simply... exceed this description character count limit. (I'll get around to it eventually)Until then, Have fun.
Anime to Realistic WF
Overview: The Ultimate Animé-to-Realistic PipelineThis asset is optimized for bridging the gap between Subculture Character Accuracy and Hyper-Realism. By decoupling geometry from texture, it allows you to cross-render any anime character into real life without the uncanny valley effect.How to use1️⃣ Phase 1: Character Geometry Lock-In* Base Processing: Generate your initial frame using Illustrious/Pony-based setups (such as WAI-Illustrious).* LoRA Injection: Load your target character LoRA at Strength 1.0.* Architect's Note: The output at this stage might look over-baked or overly stylized, but it is mandatory to anchor the exact outfit, hair, and accessory vectors.2️⃣ Phase 2: Photorealistic Coating (The ZiT Pass)Push the initial latent/image into *ZiT (CyberRealistic / CyberIllustrious)** via i2i (img2img) sampling.* Optimization: If Phase 1 yields perfect structure, you can bypass or downscale this sampling layer.3️⃣ Phase 3: Anatomical Refinement (Detailer Pass)Utilize FaceDetailer and HandDetailer node sequences to neutralize facial bias and fix micro-deformations.⚠️ Risk Management & Troubleshooting※ You may not achieve great satisfaction with some LoRAs. In this case, try modifying the existing model or adjusting the denoising.
【SDXL】MEGAPACK
✨ SDXL — Workflow MEGAPACK — Simple WorkflowHere you will find all my workflows for SDXL. The easiest way is to copy this folder to ComfyUI\user\default\workflows.✔️ What's included :TXT to IMGIMG to IMGINPAINTOUTPAINTControlNetUpscalerLoRA tester⚠️ IMPORTANT — Nodes 2.0 RequiredThis workflow is built for the Nodes 2.0 (Vue) interface of ComfyUI. If you don't enable it, the workflow may have display problems.How to activate Nodes 2.0:Open ComfyUIGo to Settings (⚙️ icon, bottom-left)Find "Use Nodes V2 (Vue)" and toggle it ONRefresh the pageLoad the workflow📦 Custom Node RequiredOnly one custom node to install:👉 ComfyUI-UmeAiRT-ToolkitInstall via ComfyUI Manager (search "UmeAiRT") or use the UmeAiRT Auto-Installer.The Toolkit packages everything internally — upscaler, face detailer, metadata saver. No other custom nodes needed.📂 Files you need An SDXL model : like this one in ComfyUI/models/checkpoints/An upscale model (optional) : like this one in ComfyUI/models/upscale_models/As many LoRAs as you like (optional) : like this one in ComfyUI/models/loras/For the Subject Detailer (optional, download automatically by my toolkit) :A BBOX detector model : like this one in ComfyUI/models/ultralytics/bbox/A SAM model : like this one in ComfyUI/models/sams/A SEGM detector model : like this one in ComfyUI/models/ultralytics/segm/
**Ultimate Illustrious Anime image and Refiner**
Ultimate Illustrious Anime Refiner – a powerful, production-grade refinement pipeline built specifically for the Illustrious ecosystem (cstatiANIME25DXL, anifoxXL, and similar anime XL models).This workflow turns good generations into stunning final artwork by fixing the biggest pain points in anime generation: deformed hands, weak eyes, inconsistent faces, and anatomy issues — while keeping beautiful 2.5D/illustrative style and excellent color/lighting.### ✨ Key Features- Smart Prompt Builder — Load any reference image → multiple WD14 Taggers (eva02, vit, swinv2, convnext) + intelligent TagFilter system that automatically separates Character / Clothing / Pose / Background tags for clean, highly detailed prompts.- Advanced Multi-Stage Detailer Chain: - Face Detailer (animeface detector) - Hand Detailer (with PerfectHandsIllustrious LoRA) - Eye Detailer (intricate iris, limbal ring, catchlights) - Nipple Detailer (final anatomy pass)- Professional Polish Stage — Iterative latent upscaling + ColorMatchV2 + RTX Video Super Resolution + brightness/contrast correction- Clean LoRA stacking system (PerfectHands, Sexy Boost v2, DetailerILv2, etc.)- Heavy VRAM optimization with cleanup nodes### Best Used For- Fixing messy or mid-quality generations- Character consistency / reference-based creation- High-end 2.5D anime illustrations- Users who want maximum hand & eye quality without manual inpainting### RequirementsCustom Nodes (must have):- ComfyUI-Impact-Pack- ComfyUI-Easy-Use- rgthree-comfy- comfyui-kjnodes- ComfyUI-Chibi-Nodes- ComfyUI-WD14-Tagger- ComfyUI LayerStyle- Mango Node Pack (optional)Recommended Models:- Main Checkpoint: cstatiANIME25DXL_v1025DSOFT.safetensors- Detailer Checkpoint: anifoxXL_anifoxV1.safetensors- LoRAs included in stacks: PerfectHandsIllustrious, sexy_boost_v2, DetailerILv2-illustrous, etc.- Upscaler: remacri_original.ptRecommended Resolution: 832×1216 or 1024×1536 (portrait)### How to Use1. Drop your reference/generated image into the LoadImage node (top left).2. The workflow automatically builds a detailed prompt using the tagger system.3. Hit Queue Prompt — everything else is fully automated.4. (Optional) RTX Super Resolution node can be bypassed if you don’t have a compatible NVIDIA card.This is currently my daily driver for anime work. Extremely consistent results with beautiful hands, eyes, and overall polish.Enjoy!
Workflow by Zar
Мои рабочие процессы
📜 DaSiWa Image Workflows | T2I | I2I 📜
My ComfyUI workflows for T2I/I2IThey are optimized for my checkpoints and created of my latest knowledge to enhance the outcome. "If this workflow leveled up your day, I'd purr-eciate a like! 😻"Versions & Information👇👇👇👇👇👇👇👇👉 Please read below and the file descriptions "About this version" for more info's.⚠️ Do not use the workflows with the "Nodes 2.0 beta" from ComfyUi or it will mess up things.👇👇👇👇👇👇👇👇What you get from the comfy workflows:♨️ Easy controls✅ As less as possible dependencies🪧 Detailed documentation⛓️ Highly automatic logic✨ Optimized results🎬 Fully automated resolution logic🔖 Bookmark-Shortcuts with number keysTypes of workflowsDarkArts C-TI2I🖼️ T2I and I2I🧩 Automatic aspect‑ratio calculation and fitting✨ Multiple Upscalers + RefinersUpscale with Model (additional detail, high quality)RTX Super Resolution (+ Deblur + Denoise)⏩ Multiple sampler passes💎 Easy Quality prompt option🎰 Multiple Radomierzers & Wildcards🔖 Bookmark-Shortcuts - with number keys🩻 Known issues and advice'snone🫶 Special thanks @TrashAI for the idea and help with the wildcards!All older Versions are available inside my GitHub Repo.YOU are responsible for outputs as always! If you make ToS violating content and I get aware I WILL report this.
Illustrious/ Pony/ SDXL SFW & NSFW Pro Grade Workflow (Low and high VRAM)
Note: I spend a lot of time on these to make sure they are top notch. If you have ANY issues, I'm very responsive. hit me up and I'll resolve them for you.This is an advanced workflow. It is node heavy and designed for multiple options and to return professional high quality results. If you are looking for simpler workflows, they are here:An essentials one can be found here: https://civitai.com/models/2349309/illustriousponysdxl-essentialsA simple one can be found here:https://civitai.red/models/2600919/lonecats-simple-zitilluanimaflux-workflows?modelVersionId=2921603A complete user and troubleshooting guide inclusing a list of common error messages can be found here https://civitai.com/articles/24678/how-to-use-my-workflows-a-comprehensive-breakdownInstagram: https://www.instagram.com/synth.studio.models/Buy me a☕ https://ko-fi.com/lonecatoneThis represents hundreds of hours of work. If you enjoy it, please 👍like, 💬 comment , and feel free to ⚡tip 😉也有中文说明VER 10.1:Got tired of having to always mess with resolution, so I gave you the option to make it automaticMath node replacements to deal with the Linux errorsAdded psst... fast bypassers (you'll thank me).💡use the upgraded face yolo I added. much more accurate.Improved notes throughoutVer 8.0/ 9.0:Major overhaul to bring it up to speed with my other workflowsAdded second checkpoint metadataadded a prompt to hi-res fixadded Inpainting suite from the pro grade workflow by request.V7.0Apparantly , I forgot to hook up Face Swap in my last model and nobody told me until today, so a free major upgrade it is!Installed a sentence remover in the prompt assist area to help eliminate CivitAI 🚩Cleaned up some bugsResinstalled Qwen VL Prompt EnhancerUpgraded Face Swapalso upgraded notes on how to remove it.Minor changes to detailersAdded an inline crop tool in the post production suite.Ver 6.0:Major overhaul to bring it up to speed with my other workflowsAdded a refiner checkpoint that can easily be switched on or offAdded FaceID as well as a style and composition IP Adapter for character consistencyReowked the Detailers ro improve thembetter version of Hi Rez fixReworked the Save with Metadata to align with CivitAi and fix the Comfy update bug 🙄This is an all in one workflow:ControlnetI2I with prompt generationDual Model Loader: Better controlSage Attention: Speeds up productionUltimate upscaler: For prescale and hi-res fixDetailers: Face, eyes, hand, expressionNSFW Detailers: Male & Female anatomySeed VR2 Upscaler: Get up to 4k pics, even on low VRAMPost production suite: My pride and joy. Does everything.
t2i-ill (Hires. Fix + Face Detailer) with spectrum nodes workflows
v1.2Added workflow descriptionReplaced the random dice nodeAdded Find & ReplaceAdded input latent variable parameters in the Hires.Fix sectionv1.1Added the link for the RES4LYF node.Replaced the WAS nodes with other equivalent nodes.The t2i-ill workflow includes the Hires. Fix workflow and the Face Detailer workflow, both of which contain Spectrum nodes.Faster image generation speed and a cleaner interface.Below are the generation speeds under different settings.If you have any better suggestions, please share them in the Discussion section.Fig.1Fig.6Fig.8
easy Inpaint workflow!
Hi!Here's a workflow for using InPaint in ComfyUI.Here's a tutorial on how to use it, although it's very easy: https://civitai.red/articles/30086¡Hola!Aquí os dejo un workflow para usar inpaint en comfyui.Aquí teneis el tutorial para usarlo, aunque es muy fácil: https://civitai.red/articles/30080
Controlnet style transfer block
Thos block is for primitive style transfer. Or, to be more acurrate, you are generating new image with desired style and trying to make it look like reference. This block is also suitable for "creative" upscale for images and restoration of blurry, low res and such images. Controlnet preprocessors are much moe omnivorous, but not almighty. Trial and error is the way here.As one of good use cases for this is to generate complex scene with bigger model (online service with qwen, z-image etc) and after use Illustrious checkpoint to make that scene look like generic or specific anime style, maybe even with characters and specific concepts, that are not possible in bigger models. Online services does not allow nsfw for free :( after all.
Sexiam - Txt2Img & Img2Img
Sexiam - Txt2Img & Img2ImgThis is not meant to be an overly complex ComfyUI workflow.The goal is to give beginner and intermediate users a flexible two-stage generation setup without overwhelming them with extra systems, heavy post-processing chains, or complicated node networks. I do a lot of Img2Img work across different styles using Illustrious-based models, and I often found myself switching between Txt2Img and Img2Img setups.I did not want separate redundant node networks doing almost the same thing, but I also did not want complicated switches that require remembering assigned numbers, bypassing groups, or breaking creative flow.The solution was to reduce the main mode switching to a simple On/Off system using boolean and if/else nodes.When the switch is off, the workflow can run as Txt2Img.When the switch is on, it can run as Img2Img.How to Use WorkflowLoad Models & Prompt1. Load Checkpoint!2. Load External VAE model3. Turn the "Use External VAE" on or off(Optional) Turn the "Use External VAE" node to true if you want to use the VAE model instead of the built-in VAE that comes with the checkpoint.4. Load your LoRA models in the "Power LoRA Loader" node.1. Press the add LoRA button.2. Select a LoRA from the list.3. To find a LoRA’s trigger words or Civitai link, right-click the LoRA in the node and select "Show Info". You can then select the trigger words you want and press copy in the "trained words" cell.5. Enter your positive and negative prompt.The positive prompt is broken up into three sections labeled Suffix, Body, and Prefix. This is to give you the flexibility to set your quality tags, LoRA trigger words, and effects tags without affecting the main prompt in the Body text node. This way you can copy and paste your character and scene tags without affecting your other modifier tags.The text encoder node network is also broken up into three separate positive prompt nodes and combined with a Condition Concat function so that checkpoints with smaller token limits have greater flexibility.KSampler - 1st Stage6. Turn the "Txt2Img" node ON if you want to use standard text-to-image processing.(Optional) Turn the "Txt2Img" node OFF if you want to use Img2Img processing instead.7. Turning ON the "Upscale by Model" switch will result in the input image being upscaled 4x by an upscaler model and then downscaled to the targeted latent space dimensions. This process can take longer than a simple latent upscale.(Optional) Turning OFF the "Upscale by Model" switch will result in the image being upscaled in latent space instead of pixel space. This is a less resource-intensive and faster process than upscale by model.8. The Resolution Master node allows you to perform a number of processes automatically when adjusting the latent space size of the generation.At the bottom of the node are a number of presets that represent most of the most-used checkpoints and their latent space sizes.In the workflow, the "Auto-Detect" option is turned on, and any image loaded into the "Load Image" node will trigger it. This will result in the node picking the closest standard size for your model.You can also upscale the generation dimensions by a factor or by total megapixels.The "Swap" button will swap the width and height.The 2D preview of the node will also show you what the orientation of the generation will look like before generation begins.9. Adjust the denoise value to your preferred setting.If you set the "Txt2Img" node to true, then the denoise setting of the KSampler will automatically be set to 1.0.If selecting Img2Img by turning the "Txt2Img" node to false, then you will need to manually adjust the setting on the pink 2D slider at the top of the KSampler - 1st Stage group.The 2D slider will adjust the integer numbers to floats automatically. For instance, if you set the slider to 70, then the KSampler will read it as 0.7 in the denoise setting.If you're only wanting small changes to the input image, then a setting lower than 50 would be recommended.If you're making large changes to the image, then a setting of 60 to 70 would be recommended.If you're curious about using Img2Img with a high denoise strength for new creative compositions, then you can refer to my article explaining the process Underpainting for AI.10. Load your preferred upscaler model and upscaling method for either upscale by model or upscale by latent.11. Adjust the KSampler settings to your preferred setup.KSampler - 2nd StageThis group will upscale the first stage and result in more detail without the need for detailer nodes.12. Adjust the denoise value to a lower setting for the upscale to work.A setting over 40 isn't recommended on most Illustrious checkpoints since you will get big changes to the upscaled image.To avoid changes to eyes, details, and composition, a denoise setting below 40 is recommended.13. Adjust the "Upscale Image By" node to your preferred upscale factor.A setting of 2 means the image size will be doubled, and 1.5 means 1.5x or 50% bigger.Depending on how much VRAM you have, you may need to adjust this to a lower value like 1.5.14. Load your preferred upscaler model and upscaling method.15. Adjust the settings for the second-stage KSampler to your preferred settings.Generally, you can set the CFG of the second stage to a lower setting than you did with the first stage.Setting the CFG too low may result in a color shift, though.16. The "SD Prompt Saver" is set up to organize your generations by Model Name folder.The saved images should also work with Civitai's system for metadata retrieval.
Controlnet inpainting block
To create my videos I do use a loong chain of inpaintings. This workflow is just a building block for such chain. There is nothing new, but feel free to get ideas from it.
Make your own yugioh cards!
Workflows for yugioh cards.The font should go in: Documents\ComfyUI\custom_nodes\ComfyUI_Comfyroll_CustomNodes\fontsAquí os dejo 1 workflow para cartas de Yugioh.La fuente deber ir en: Documents\ComfyUI\custom_nodes\ComfyUI_Comfyroll_CustomNodes\fonts
Make your own Digimon Card!
Hi! With this workflow you can make your own Digimon card!The font should go in: Documents\ComfyUI\custom_nodes\ComfyUI_Comfyroll_CustomNodes\fontsAnd to change the text, go into the subgraph.¡Hola! ¡Con este workflow puedes hacer tu propia carta de Digimon!La fuente deber ir en: Documents\ComfyUI\custom_nodes\ComfyUI_Comfyroll_CustomNodes\fontsY para cambiar el texto meteos en el subgrafo.
Make your own visa cards!
Hi! With this workflow you can create your own Visa card images.¡Hola! Con este workflow podrás hacer tus propias imágenes de tarjetas visa.
ComfyUI Image Workflows
Workflow for XL/Illustrious/NoobAI ModelsAny feedback would be appreciated.See the Article for a Guide:https://civitai.com/articles/17339 (probably doesn't work rn)https://civitai.red/articles/17339Key Features:LoRATIPO PromptingDetail EnhancersControlNet UpscaleIPAdapterHiresFixOpenPoseWildcardsControlNETColor MatchFBCNNRegional PromptingInpaintOutpaintWorkflows included:🟥Advanced: comprehensive text to image workflow🟨Basic: more basic version of the Standard workflow🟦Standard: more basic version of the Advanced workflow🟩Detailer: detailing previously generated or existing imagesExperiment and enjoy!
Metadata Transfer v2
Metadata Transfer v2This workflow lets you copy readable generation metadata from one image and write it to another image. It is designed for cases where an image still looks correct, but the embedded metadata is missing, incomplete, or no longer matches your current setup.This is especially useful when preparing images for Civitai, recovering metadata after external editing, or correcting details after renaming models or changing file paths. Instead of manually rebuilding the settings, the workflow can pull the available data from a source image and apply it to an identical target image, including key fields such as the model name and KSampler settings.The goal is simple: keep your image metadata clean, accurate, and ready for upload without having to recreate the generation info by hand.How to Use:Load the image that contains the correct metadata into the “SD Prompt Reader” node. This is your source image.Load the image you want to update into the “Load Image” node. This is the target image that will receive the metadata.Enable “Manually Input Settings” only if you want to override the automatic metadata transfer and enter the values yourself. Leave it off if you want the workflow to pull the settings directly from the source image.If the source image used any LoRAs, load those LoRA models into the “Wildcard Encode (Inspire)” node. This allows the workflow to include the LoRA information when writing the target image’s metadata.Note:When “Manually Input Settings” is disabled, the workflow will automatically read the available settings from the source image and write them into the target image’s metadata. LoRAs are the one exception: they must be loaded manually in “Wildcard Encode (Inspire)” if you want them included, even when automatic transfer is enabled.Manual InputIf you want to enter the metadata yourself, first enable the “Manually Input Settings” boolean node. This tells the workflow to use the values you provide manually instead of reading them automatically from the source image.Toggle “Manually Input Settings” on.Load the checkpoint/model you want written into the metadata.Enter the KSampler settings, including the sampler, scheduler, steps, CFG scale, and seed.Enter the positive prompt.Enter the negative prompt.Optional: If the image used LoRAs, load the LoRA models into the “Wildcard Encode (Inspire)” node so they can be included in the final metadata. After the manual values are entered, run the workflow. The target image will be saved with the metadata values you provided.
Extremely Simplified SDXL Workflow APP
Extremely Simplified SDXL WorkflowA clean, no-nonsense SDXL workflow built for people who want results without fighting their node graph. Whether you are brand new to ComfyUI or just tired of bloated workflows, this gets you generating fast. Everything that matters is surfaced in APP Mode - prompts, model, sampler settings, resolution, and LoRAs. Everything else stays out of your way.Who Is This For?- ComfyUI beginners who want a solid starting point without being overwhelmed- No-nonsense users who just want to hit generate and get a good image- Low VRAM users - aggressive GPU memory cleanup nodes are placed throughout the workflow to keep things stable- If you have ever opened a "simple" workflow and found 200 nodes staring back at you, this is the antidoteVRAM Requirements| VRAM | Experience ||---|---|| 8 GB | Suitable - cleanup nodes help keep it stable || 12 GB | Recommended sweet spot || 16 GB+ | Smooth, no compromises |The workflow uses dedicated VRAM cleanup nodes between passes so even lower-end cards can get through the full two-pass pipeline without running out of memory.What's Included- Two-pass generation with a built-in hires fix (first pass + upscale/refine)- rgthree KSampler Config - set steps, CFG, sampler, and scheduler once, applies to both passes automatically- Lora Loader Stack with 4 LoRA slots- SAM + bbox detector for the second pass (automatic subject detection and detailing)- Resolution Selector with proper SDXL aspect ratios and megapixel scaling built in- CLIPSetLastLayer at -2 for cleaner conditioning- GPU memory cleanup nodes between passes so VRAM does not get eaten alive- You only see the final result - the first pass preview is intentional and hidden to keep things clean- The detailers come pre-tuned with balanced settings that should work well for the vast majority of users. If you find the results are not quite to your liking, you will need to switch to workflow mode and adjust the detailer nodes manually - that said, the defaults I chose should be more than sufficient for 90% of use cases.The Work Behind ThisThis workflow went through nearly 4 hours of testing, tweaking, and iteration to find the right balance between simplicity and output quality. Every node placement, every cleanup point, and every default setting was chosen deliberately. The goal was to build something that a beginner can pick up instantly but that does not cut corners on what matters.Required Custom NodesInstall these via ComfyUI Manager before loading the workflow:- rgthree-comfy- ComfyUI Easy Use- ComfyUI Impact PackRequired Modelsbbox Detector (for the detailer/second pass)The workflow uses Eyeful_v2-Paired.pt as the default bbox detector. This is NOT included with Impact Pack by default.Download it here:👉 Eyeful_v2_pairedInstallation:1. Download Eyeful_v2-Paired.pt (or any other .pt bbox model you prefer)2. Place it in: ComfyUI/models/ultralytics/bbox/3. Create the folder if it does not exist4. Restart ComfyUI - it will appear in the model dropdown automatically> You can swap it for any other bbox model from the adetailer collection such as face_yolov8m.pt or hand_yolov8s.pt - same install location.Default Settings| Setting | Value ||---|---|| First Pass Steps | 30 || Hires Fix Steps | 15 || CFG | 4.5 || Sampler | Euler || Scheduler | Simple || Resolution | 1344x768 (16:9 @ 1MP) |Change any of it from the Linear Mode panel without touching the graph.Designed for people who know what they want to generate, not people who want to spend 2 hours wiring nodes.
Arthemy Merge - ComfyUI
✦ Arthemy Merge Model/Clip ✦INSTRUCTIONS & DOCUMENTATIONWelcome to the Arthemy SDXL workflow!Standard merging often "flattens" the AI landscape because the resulting model just slides somewhere in between two extremes.I wanted a way to surgically slice, tune, and repair my models (and their CLIPs) to get the exact aesthetic I had in mind. Since native ComfyUI nodes couldn't do exactly what I wanted, I spent hours coding and debugging this custom suite (with the help of an LLM, because I'm a noob... but they work!).This workflow is designed to let you mix in memory, test live, and bake later. Here is a breakdown of how the workflow is structured, what the custom nodes do, and how you can use the values to cook your own ultimate model.InstallationUnzip the fileMove the "Arthemy_SDXL_Suite.py" in your "[...]ComfyUI/custom_nodes/" folder.Drag and Drop the workflow in your ComfyUI window and you're ready to go!📦 The Setup: Live Testing & Saving SafelyThe most effective way to do this: mix models in memory, test the combination live, and only save the version that actually improves your output.The Loaders: Load your baked Checkpoint or model, CLIP, and VAE separately. Pro-tip: in the later stages of your workflow, it might be useful to experiment with how different iterations of your CLIP and Model behave between them.The Live Test: This is for your quick-and-dirty test generations. Keep your Seed on 'Fixed' and pick your target prompt. Copy this entire group to test your Model and CLIP at different stages of the process, or to try out different settings and prompts on the fly.These are the extremes of the workflow you’re going to assemble. Between them, you can stack as many of the following nodes as you want. When the output looks good, you can just wire the output to the “Save Model / CLIP” group and “bake” all of these changes into a new model.💾 The Arthemy Checkpoint SaverWhen the output looks good, you can wire the final MODEL, CLIP, and VAE outputs to this node to "bake" all of your changes into a brand-new model.What it does: Standard ComfyUI save nodes often struggle to save modified CLIPs or fail to capture live memory patches properly. This custom Saver ignores ComfyUI's memory manager, extracts the "cooked" weights one by one, maps the SDXL dual-encoder keys perfectly to the Vanilla format, and streams them to your hard drive.Why it's better: It guarantees that your output .safetensors file is 100% compatible with Civitai, A1111, Forge, and any standard loader.LoRA Loaders (Model & CLIP)This is one of the simplest groups.You can use it to inject your LoRAs into a model, baking them into it. Since some LoRA with a very limited scope might ruin your model's flexibility, I highly suggest to use this group with LoRA trained for a Style (with a very wide scope).To use this, I generally test a lot of LoRAs, one at the time, with extreme positive and negative values in both the Model and the CLIP department (just to see how they can be used in any scenario).Then, I stack all of these LoRAs in the group and I start to tweak their weights in order to make them influence the model without being too destructive. In general I suggest to move between "-0,2" and "0.2".Simple Merge (Model & CLIP)The simplest way to change your model is to merge it with a secondary model.1.0 -> 100% Your base Model or CLIP.0.5 -> 50% Your base Model or CLIP / 50% Model or CLIP 2.0.0 -> 100% Model or CLIP 2.I know this, you know this, everybody knows this, but you can also use a model as an external ingredient by using weights ABOVE 1.0.1.2 -> Your base Model or CLIP is now merging an inverse version of the other model, moving mathematically away from its values.🧬 The Block Mergers (Model & CLIP)What it does: Instead of blending whole models, you can slice them into semantic parts and choose how each slice is affected by the merge individually.How the values work:1.0 -> 100% Your Base Model.0.5 -> 50% Your Base / 50% Secondary Model.0.0 -> 100% Secondary Model.Pro-Move: You aren't locked between 0 and 1! You can go above 1.0 or below 0.0 to do an inverse-merge, mathematically pushing your base model away from the secondary model's style.✨ The Tuners (Model & CLIP)What it does: You don't need a second model for this. The Tuner works as a Multiplier: it scales the internal math of your model up or down to boost or fade specific visual concepts (like Textures or Composition) or CLIP concepts (like Semantic Focus).How the values work:1.0 -> 100% (Leaves the concept exactly as it is).0.8 -> 80% (Fades the concept, softening its intensity by 20%).1.2 -> 120% (Boosts the concept, amplifying its intensity by 20%).Pro-Move: Go way past these numbers (e.g., 2.0 or -0.5) to violently force a style or completely mute a concept out of existence.🎲 The Chaos Block Mergers (Model & CLIP)What it does: Sometimes you don't know exactly what you want, but you want to discover a completely new aesthetic. The Chaos Merger automates the Block Merge process by assigning randomized weights to every single semantic block of the model.How the values work: You set a "Chaos Seed". For every block in the model (Input, Middle, Output, or CLIP layers), the node generates a random blend ratio based on that seed.Low Chaos (e.g., 0.1 to 0.3): The random weights stay very close to your Base Model (mostly 1.0). It adds subtle, unpredictable variations to the style.High Chaos (e.g., 0.8 to 1.5): The weights swing wildly between your Base Model, the Secondary Model, and even extreme inverse-merging values. This creates chaotic, entirely new visual identities.Found a crazy combination you love? Save the Seed number! The math is deterministic; using the same seed and the same models will always yield the exact same random block weights.🌪️ The Chaos Block Tuners (Model & CLIP)What it does: Similar to the Chaos Merger, but it applies randomization to the Tuner multipliers instead of blending two models. It randomly amplifies or dampens the internal math of your Base Model without needing any secondary ingredients.How the values work: Set a Chaos Seed and a Chaos amount.Low Chaos: The internal block multipliers flutter slightly around 1.0 (e.g., between 0.9 and 1.1). Great for adding a touch of unpredictability to a stagnant model.High Chaos: The multipliers swing wildly (e.g., boosting a block to 1.8 while crushing another to 0.2). This can totally break a model—or give it a completely unique, hyper-stylized look.Because you aren't introducing new concepts from a second model, the Chaos Tuner is the ultimate tool for discovering hidden aesthetics already inside your current checkpoint.🔧 The Restorers (Model & CLIP)What it does: This is your Repair Tool. When you push a model or a CLIP too far with merges, it gets "fried" (deep-fried contrast, weird color hazes). Unlike the Tuner, the Restorer acts as a Ceiling and a Subtractor. It irons out the broken math without ruining healthy contrast, attacking only the extreme numbers.1. Spike Flattening (Variance Ceiling): Use this if your model has burned, overly harsh contrast.99.0 -> Off (Leaves all the spikes and extreme values untouched).2.5 -> Mild Flattening (Tames only the worst, most broken spikes).1.5 -> Aggressive Flattening (Heavily flattens the overall contrast).2. Offset Flattening (Centering the Drift): Use this if your model is spitting out weird color tints or a foggy baseline.0.0 -> 0% Flattening (Leaves the model's baseline exactly as it is).0.5 -> 50% Flattening (Gently centers the wandering numbers).1.0 -> 100% Flattening (Completely neutralizes the offset, pulling the math back to a healthy zero).Pro-Move: You can go above these boundaries to over-correct a deeply fried model, or play with them to intentionally distort a healthy one.This workflow is Modular!You can create any combinations of this groups and make this workflow work in the way you prefere. Have fun breaking things, testing limits, and building the exact tools you need to get the images you want!
简单易懂的Illustrious工作流/text2img：Simple Workflow for Illustrious(lora、fix、FaceDetailer)
独创的三步生图法，目的是解决lora、画师tag等对模型生图能力固化的现象，通过多步生成赋予最大的生成操作空间。这个工作流逻辑简单易懂，包含lora、放大、面部修复，一站式生图。默认模型仅供测试使用。您可以使用任何您喜欢的模型。面部检测模型下载地址Anzhc Face seg 640 v4 y11n.pt · Anzhc/Anzhcs_YOLOs at main眼部检测模型下载地址Anzhc Eyes -seg-hd.pt · Anzhc/Anzhcs_YOLOs at main4x-UltraSharpV2_Lite模型下载地址https://huggingface.co/Kim2091/UltraSharpV2/tree/main线条ControlNet下载地址mistoLine_rank256.safetensors · TheMistoAI/MistoLine at mainThis workflow logic is relatively simple, but it includes lora, amplification, and repair.The default model is only for testing purposes. You can use any model you like.Face detection model download address Anzhc Face seg 640 v4 y11n.pt · Anzhc/Anzhcs_YOLOs at mainEye detection model download address Anzhc Eyes -seg-hd.pt · Anzhc/Anzhcs_YOLOs at main4 x - UltraSharpV2_Lite model download address https://huggingface.co/Kim2091/UltraSharpV2/tree/mainControlNet Line model download address mistoLine_rank256.safetensors · TheMistoAI/MistoLine at main
Simple SDXL workflow with upscaling RTX | SVR2
A simple Comfyui Workflow I use for basic testing and image generation. Contains the "RTX Video Super Resolution" upscaling node to upscale images and videos. Only works on Nvidia RTX GPUs. https://github.com/Comfy-Org/Nvidia_RTX_Nodes_ComfyUI.git05/2/26 updateAlso added seed vr2 version for non nvidia users. The quality is about the same though I prefer RTX solution as it is notably faster (almost instant) and I prefer the upscalling quality slightly better. Still probably better to go with a good forge workflow :p.
Make your own ninja charts in the style of Naruto!!
Hi!I've created this simple workflow so you can make a ninja character sheet for any character. You can choose any village you want and even add the name right there in the workflow.I've included a character sheet with civitai logo, a logo I made for my character's village and a logo for Albert Wesker's village, assuming he's in the Naruto universe. Grok helped me with it, haha.I hope you like it!!https://civitai.red/articles/29371/how-to-use-the-cards-and-ninja-charts-workflows¡Hola!He hecho este simple workflow para poder hacer la ficha ninja de cualquier personaje, podéis escoger la aldea que queráis e incluso ponerle el nombre en el mismo workflow.He incluido una ficha con el logo de civitai, otra con un logo que he hecho de la aldea de mi personaje y otra con la de la aldea de Albert Wesker en el supuesto de que estuviera en el universo de Naruto. Grok me ayudó a hacerlo jeje.¡¡Espero que os guste!!https://civitai.red/articles/29370You should place the text fonts here: C:\Users\etooo\Documents\ComfyUI\custom_nodes\ComfyUI_Comfyroll_CustomNodes\fontsSimply use the play button on the node to see how the text will look.Click on the node called NAME PREVIEW and the blue play button will appear; click it to see how the text will look.Las fuentes de texto debéis ponerlas aquí: C:\Users\etooo\Documents\ComfyUI\custom_nodes\ComfyUI_Comfyroll_CustomNodes\fontsSolo debéis usar el play en el nodo para ver cómo quedará el texto.Hacemos click en el nodo llamado NAME PREVIEW y aparecer el botoncito azul del play, le damos para ver cómo quedaría el texto.Now we're going to change the character's name.Ahora vamos a cambiarle el nombre al personaje.We'll use the little blue play button again to see how it would look with the change.Volvemos a usar el botoncito de play azul para ver cómo quedaría con el cambio.Since it's too small, we're going to resize it:Como queda demasiado pequeño, le vamos a cambiar el tamaño:Let's go back and see how it would look:Volvemos a darle a ver cómo quedaría:Now it's a little misaligned, so we're going to move it so that it's properly centered:Ahora se ha quedado un poco descuadrado, así que lo vamos a mover para que esté bien centrado:We press the blue play button again:Volvemos a darle al botoncito azul de play:Once it's perfect, we run the entire workflow.Cuando ya esté perfecto le damos a ejecutar todo el workflow.
Make your own custom Love Live idolized icon!!
¡Hola!Es muy fácil de usar, simplemente eliges el icono y a tu waifu y su expresión y le das a ejecutar.Hi!It's very easy to use, you simply choose the icon, your waifu and her expression, and then press run.
Ultimate ComfyUI SDXL/PDXL/Illustrious Workflow
Ultimate ComfyUI SDXL/PDXL/Illustrious WorkflowMy ComfyUI SDXL/PDXL/Illustrious workflow features three upscale/refine stages, OpenPose ControlNet, numerous detailers, and a neatly organized structure. Most options are easily configurable via the dashboard, using dropdowns or boolean toggles.Notes:You will need to download the various bbox models to successfully run the detailers. You can find these models on huggingface/civitai. You can find the SAM model in the ComfyUI model manager.You can disable detailers by bypassing the detailer node. You can also disable the entire group and the model loader. However, you must leave the get_DETAILER_*_IMAGE nodes enabled.Features:- Dual Model Setup: Enables a stylized model for initial generation and a secondary model for refinement/upscale stages. I typically use a PDXL/Illustrious model as the secondary one, as it excels at handling human anatomy, hands, and faces.- Three-Stage Sample/Upscale Design: Provides full control over upscaling and sampling, unlike consolidated HiRes Fix nodes. The incremental approach reduces artifacts and distortions while preserving details from the original generation.- Control Panel: Centralizes easy configuration of models, settings, and more—all in one place.- Consolidated Prompts: Dedicated locations for all prompts and detailer prompts.- Prompt Bypass Options: Bypass stage 2/3 prompts with a single click to use one positive/negative prompt across all stages. Selectively bypass positive, negative, or both.- Full Scaling Control: Adjust scaling factors for each stage independently.- Pony Toggles: Automatically add Pony scores to prompts if desired, keeping prompts clean and organized. This can be edited, and the text replaced, for embeddings instead.- Image Loader and Switch: Supports img2img or txt2img modes, with optional inversion and tinting.- LoRA Stacks: Separate stacks for stage 1, stages 2/3, and detailers, offering maximum control over LoRA application.- Automatic Image Dimensions: Handles width and height automatically (conditioning and scaling).- Eight Detailers: Includes hands, breasts, penis, butt, vagina, face, eyes/mouth, and jewelry.- Precise Masking: All detailers use bounding box (bbox) and Segment Anything Model (SAM) for accurate masks.- Hand Detailer: Utilizes OpenPose ControlNets to guide refinements for hands and fingers.- Face Detailer: Employs MediaPipe FaceMesh for ultimate control, detecting face, eyes, pupils, mouth, etc., individually or together.Dependencies (Custom Nodes):Comfyroll StudioCustom ScriptsDream ProjectImpact PackInspire PackKJNodesTiled DiffusionEasy Use
Simple Synthartx workflow
It takes around 80 seconds for a generation using the LOW_VRAM configuration on a RTX 5080(16 GB VRAM), 64 GB OF RAMIt takes around 380 seconds for a generation using the HIGH_VRAM configuration on a RTX 5080(16 GB VRAM), 64 GB OF RAMLoad any checkpoint you want, but I used janku 7.77 for my generations. Please note that different models have different cfgs, samplers, and schedulers. I've left them easily configurable in the processing suit.You are free to use my set of loras, but if you have your own than you can change them in the lora manager.My set of Loras -<lora:AddMicroDetails_Illustrious_v5:0.35>, <lora:Dramatic Lighting Slider:1.00>, <lora:outline_xl_end:-1.00>, <lora:MOON-NSFW:0.40>, <lora:E7BBAAE584BF20E58E9AE6B682E69E84E59BBE.gkED:0.60>, <lora:Rella-000195:0.80>My newest set of loras -lora:AddMicroDetails_Illustrious_v6:0.35> <lora:MOON-NSFW:0.40> <lora:E7BBAAE584BF20E58E9AE6B682E69E84E59BBE.gkED:0.25> <lora:748cm_c_illu:0.30> <lora:USNR_STYLE_ILL_V1_lokr3-000024:0.80> <lora:Dramatic Lighting Slider:1.00> <lora:Rella-000195:0.60> <lora:Create Concept (Style) Illustrious:0.40> <lora:style_strength_controller_illus01_v1:0.40>E7BBAAE584BF20E58E9AE6B682E69E84E59BBE.gkED -> [Style filter] Xu Er thick paint composition light texture enhancementIn the Load Model SubgraphYou can change the Width, Height, and Batch Size (how many generations at once)Your positive prompt is the top boxYour negative prompt is the bottom boxKS Pass 1 is Ksampler 1-60 steps recommended - 15 steps minimumdpmpp_3m_sde_gpu recommended - Euler A minimumkarras recommended - simple minimumDetailer120 steps recommended - 15 steps minimumdpmpp_3m_sde_gpu recommended - Euler A minimumkarras recommended - simple minimumIterative Upscale3 iterative steps recommended - 2 iterative steps minimumdpmpp_3m_sde_gpu recommended - Euler minimumkarras recommended - simple minimum3x upscale factor recommended - 1.5x minimum10 steps recommended - 8 steps minimum0.35 denoise recommended - 0.2 minimumYou are free to choose what image saver you want, I just put the basic one.Extra notes-Iterative upscale is quite heavy, so you can just bypass it if your computer can't handle it.The results of Kspass 1 and detailer are already good if you can set that higher.Thank you for using my workflow, ask me for any questions
【SDXL】ControlNet
✨ SDXL — ControlNet — Simple WorkflowA clean, all-in-one SDXL ControlNet workflow built entirely with the UmeAiRT Toolkit for ComfyUI.This workflow allows you to generate images from text inspired by an existing image using a ControlNet processor.Only 8 nodes. No spaghetti wires. Just load your model, write your prompt, and hit generate.You can find a step-by-step guide to this workflow in this article: link⚠️ IMPORTANT — Nodes 2.0 RequiredThis workflow is built for the Nodes 2.0 (Vue) interface of ComfyUI. If you don't enable it, the workflow may have display problems.How to activate Nodes 2.0:Open ComfyUIGo to Settings (⚙️ icon, bottom-left)Find "Use Nodes V2 (Vue)" and toggle it ONRefresh the pageLoad the workflowIf you prefer the classic interface, check out my Legacy version of this workflow instead (link).🎯 FeaturesText-to-Image generation with any SDXL/ILLUSTRIOUS checkpointBuilt-in upscaler — high-quality tiled upscaling (toggleable on/off) Slower than a classic upscaler, but significantly better qualitySubject Detailer — automatic face or body enhancement (toggleable on/off)Full metadata embedding — your images are saved with all generation parameters, ready for online publishing and remixing3 LoRA slots — with individual on/off toggles and strength control and you can connect as many other lora modules to each other for as many LoRA as you want.📦 Custom Node RequiredOnly one custom node to install:👉 ComfyUI-UmeAiRT-ToolkitInstall via ComfyUI Manager (search "UmeAiRT") or use the UmeAiRT Auto-Installer.The Toolkit packages everything internally — upscaler, face detailer, metadata saver. No other custom nodes needed.📂 Files you needAn SDXL model : like this one in ComfyUI/models/checkpoints/An upscale model (optional) : like this one in ComfyUI/models/upscale_models/As many LoRAs as you like (optional) : like this one in ComfyUI/models/loras/For the Subject Detailer (optional, download automatically by my toolkit) :A BBOX detector model : like this one in ComfyUI/models/ultralytics/bbox/A SAM model : like this one in ComfyUI/models/sams/A SEGM detector model : like this one in ComfyUI/models/ultralytics/segm/
NoobAI/Illustrious Workflow
Just a simple ComfyUI workflow for NoobAI/Illustrious
【SDXL】LoRA TESTER
✨ SDXL — LoRA Tester — Simple WorkflowA clean, all-in-one SDXL LoRA Tester workflow built entirely with the UmeAiRT Toolkit for ComfyUI.Only 8 nodes. No spaghetti wires. Just load your model, write your prompt, and hit generate.⚠️ IMPORTANT — Nodes 2.0 RequiredThis workflow is built for the Nodes 2.0 (Vue) interface of ComfyUI. If you don't enable it, the workflow may have display problems.How to activate Nodes 2.0:Open ComfyUIGo to Settings (⚙️ icon, bottom-left)Find "Use Nodes V2 (Vue)" and toggle it ONRefresh the pageLoad the workflowIf you prefer the classic interface, check out my Legacy version of this workflow instead (link).🎯 FeaturesText-to-Image generation with any SDXL/ILLUSTRIOUS/PONY checkpointBuilt-in upscaler — high-quality tiled upscaling (toggleable on/off) Slower than a classic upscaler, but significantly better qualitySubject Detailer — automatic face or body enhancement (toggleable on/off)Full metadata embedding — your images are saved with all generation parameters, ready for online publishing and remixing3 LoRA slots — with individual on/off toggles and strength control and you can connect as many other lora modules to each other for as many LoRA as you want.📦 Custom Node RequiredOnly one custom node to install:👉 ComfyUI-UmeAiRT-ToolkitInstall via ComfyUI Manager (search "UmeAiRT") or use the UmeAiRT Auto-Installer.The Toolkit packages everything internally — upscaler, face detailer, metadata saver. No other custom nodes needed.📂 Files you needAn SDXL model : like this one in ComfyUI/models/checkpoints/An upscale model (optional) : like this one in ComfyUI/models/upscale_models/As many LoRAs as you like (optional) : like this one in ComfyUI/models/loras/
Afnm Workflows
Workflows to produce assets for the game Ascend from Nine Mountains
【SDXL】OUTPAINT
✨ SDXL — Outpaint — Simple WorkflowA clean, all-in-one SDXL outpaint workflow built entirely with the UmeAiRT Toolkit for ComfyUI.Only 8 nodes. No spaghetti wires. Just load your model, write your prompt, and hit generate.You can find a step-by-step guide to this workflow in this article: link⚠️ IMPORTANT — Nodes 2.0 RequiredThis workflow is built for the Nodes 2.0 (Vue) interface of ComfyUI. If you don't enable it, the workflow may have display problems.How to activate Nodes 2.0:Open ComfyUIGo to Settings (⚙️ icon, bottom-left)Find "Use Nodes V2 (Vue)" and toggle it ONRefresh the pageLoad the workflowIf you prefer the classic interface, check out my Legacy version of this workflow instead (link).🎯 FeaturesText-to-Image generation with any SDXL/ILLUSTRIOUS/PONY checkpointBuilt-in upscaler — high-quality tiled upscaling (toggleable on/off) Slower than a classic upscaler, but significantly better qualitySubject Detailer — automatic face or body enhancement (toggleable on/off)Full metadata embedding — your images are saved with all generation parameters, ready for online publishing and remixing3 LoRA slots — with individual on/off toggles and strength control and you can connect as many other lora modules to each other for as many LoRA as you want.📦 Custom Node RequiredOnly one custom node to install:👉 ComfyUI-UmeAiRT-ToolkitInstall via ComfyUI Manager (search "UmeAiRT") or use the UmeAiRT Auto-Installer.The Toolkit packages everything internally — upscaler, face detailer, metadata saver. No other custom nodes needed.📂 Files you needAn SDXL model : like this one in ComfyUI/models/checkpoints/An upscale model (optional) : like this one in ComfyUI/models/upscale_models/As many LoRAs as you like (optional) : like this one in ComfyUI/models/loras/For the Subject Detailer (optional, download automatically by my toolkit) :A BBOX detector model : like this one in ComfyUI/models/ultralytics/bbox/A SAM model : like this one in ComfyUI/models/sams/A SEGM detector model : like this one in ComfyUI/models/ultralytics/segm/
Compact and simple workflow
Compact and simple workflow designed for beginners, with the essentials needed to get started.Uses nine nodes (including external nodes) to avoid visual overload and facilitate learning.Flujo de trabajo compacto y sencillo diseñado para principiantes, con lo esencial para empezar.Utiliza nueve nodos (incluidos nodos externos) para evitar la sobrecarga visual y facilitar el aprendizaje.
Workflow backup
Workflow backup in November 2025.If you don't know what this is, don't download it.
CyberRealistic Illustrious ComfyUI Workflow
Join The Tinkerer on Whop to get early releases, private pages, and the Tinkerer Discord role - all in one place. 👉 Join on WhopFuture updates and beta versions are available via the public Whop page:https://whop.com/cyberdelia-ai-lab/comfy-ui-workflows-illustrious-d6z3nag8NuPXGE/app/Custom nodes:ComfyUI Image Metadata Extensionhttps://github.com/cyberdeliaAI/revived_comfyui_image_metadata_extensionComfyUI-PromptFormathttps://github.com/cyberdeliaAI/ComfyUI-PromptFormat
【SDXL】INPAINT
✨ SDXL — Inpaint — Simple WorkflowA clean, all-in-one SDXL inpaint workflow built entirely with the UmeAiRT Toolkit for ComfyUI.Only 8 nodes. No spaghetti wires. Just load your model, write your prompt, and hit generate.You can find a step-by-step guide to this workflow in this article: link⚠️ IMPORTANT — Nodes 2.0 RequiredThis workflow is built for the Nodes 2.0 (Vue) interface of ComfyUI. If you don't enable it, the workflow may have display problems.How to activate Nodes 2.0:Open ComfyUIGo to Settings (⚙️ icon, bottom-left)Find "Use Nodes V2 (Vue)" and toggle it ONRefresh the pageLoad the workflowIf you prefer the classic interface, check out my Legacy version of this workflow instead (link).🎯 FeaturesInpaint generation with any SDXL/ILLUSTRIOUS/PONY checkpointBuilt-in upscaler — high-quality tiled upscaling (toggleable on/off) Slower than a classic upscaler, but significantly better qualitySubject Detailer — automatic face or body enhancement (toggleable on/off)Full metadata embedding — your images are saved with all generation parameters, ready for online publishing and remixing3 LoRA slots — with individual on/off toggles and strength control and you can connect as many other lora modules to each other for as many LoRA as you want.📦 Custom Node RequiredOnly one custom node to install:👉 ComfyUI-UmeAiRT-ToolkitInstall via ComfyUI Manager (search "UmeAiRT") or use the UmeAiRT Auto-Installer.The Toolkit packages everything internally — upscaler, face detailer, metadata saver. No other custom nodes needed.📂 Files you needAn SDXL model : like this one in ComfyUI/models/checkpoints/An upscale model (optional) : like this one in ComfyUI/models/upscale_models/As many LoRAs as you like (optional) : like this one in ComfyUI/models/loras/For the Subject Detailer (optional, download automatically by my toolkit) :A BBOX detector model : like this one in ComfyUI/models/ultralytics/bbox/A SAM model : like this one in ComfyUI/models/sams/A SEGM detector model : like this one in ComfyUI/models/ultralytics/segm/
【SDXL】IMG to IMG
✨ SDXL — Image to Image — Simple WorkflowA clean, all-in-one SDXL image-to-image workflow built entirely with the UmeAiRT Toolkit for ComfyUI.Only 8 nodes. No spaghetti wires. Just load your model, write your prompt, and hit generate.You can find a step-by-step guide to this workflow in this article: link⚠️ IMPORTANT — Nodes 2.0 RequiredThis workflow is built for the Nodes 2.0 (Vue) interface of ComfyUI. If you don't enable it, the workflow may have display problems.How to activate Nodes 2.0:Open ComfyUIGo to Settings (⚙️ icon, bottom-left)Find "Use Nodes V2 (Vue)" and toggle it ONRefresh the pageLoad the workflowIf you prefer the classic interface, check out my Legacy version of this workflow instead (link).🎯 FeaturesText-to-Image generation with any SDXL/ILLUSTRIOUS checkpointBuilt-in upscaler — high-quality tiled upscaling (toggleable on/off) Slower than a classic upscaler, but significantly better qualitySubject Detailer — automatic face or body enhancement (toggleable on/off)Full metadata embedding — your images are saved with all generation parameters, ready for online publishing and remixing3 LoRA slots — with individual on/off toggles and strength control and you can connect as many other lora modules to each other for as many LoRA as you want.📦 Custom Node RequiredOnly one custom node to install:👉 ComfyUI-UmeAiRT-ToolkitInstall via ComfyUI Manager (search "UmeAiRT") or use the UmeAiRT Auto-Installer.The Toolkit packages everything internally — upscaler, face detailer, metadata saver. No other custom nodes needed.📂 Files you needAn SDXL model : like this one in ComfyUI/models/checkpoints/An upscale model (optional) : like this one in ComfyUI/models/upscale_models/As many LoRAs as you like (optional) : like this one in ComfyUI/models/loras/For the Subject Detailer (optional, download automatically by my toolkit) :A BBOX detector model : like this one in ComfyUI/models/ultralytics/bbox/A SAM model : like this one in ComfyUI/models/sams/A SEGM detector model : like this one in ComfyUI/models/ultralytics/segm/
【SDXL】TXT to IMG
✨ SDXL — Text to Image — Simple WorkflowA clean, all-in-one SDXL text-to-image workflow built entirely with the UmeAiRT Toolkit for ComfyUI.Only 8 nodes. No spaghetti wires. Just load your model, write your prompt, and hit generate.You can find a step-by-step guide to this workflow in this article: link⚠️ IMPORTANT — Nodes 2.0 RequiredThis workflow is built for the Nodes 2.0 (Vue) interface of ComfyUI. If you don't enable it, the workflow may have display problems.How to activate Nodes 2.0:Open ComfyUIGo to Settings (⚙️ icon, bottom-left)Find "Use Nodes V2 (Vue)" and toggle it ONRefresh the pageLoad the workflowIf you prefer the classic interface, check out my Legacy version of this workflow instead (link).🎯 FeaturesText-to-Image generation with any SDXL/ILLUSTRIOUS/PONY checkpointBuilt-in upscaler — high-quality tiled upscaling (toggleable on/off) Slower than a classic upscaler, but significantly better qualitySubject Detailer — automatic face or body enhancement (toggleable on/off)Full metadata embedding — your images are saved with all generation parameters, ready for online publishing and remixing3 LoRA slots — with individual on/off toggles and strength control and you can connect as many other lora modules to each other for as many LoRA as you want.📦 Custom Node RequiredOnly one custom node to install:👉 ComfyUI-UmeAiRT-ToolkitInstall via ComfyUI Manager (search "UmeAiRT") or use the UmeAiRT Auto-Installer.The Toolkit packages everything internally — upscaler, face detailer, metadata saver. No other custom nodes needed.📂 Files you needAn SDXL model : like this one in ComfyUI/models/checkpoints/An upscale model (optional) : like this one in ComfyUI/models/upscale_models/As many LoRAs as you like (optional) : like this one in ComfyUI/models/loras/For the Subject Detailer (optional, download automatically by my toolkit) :A BBOX detector model : like this one in ComfyUI/models/ultralytics/bbox/A SAM model : like this one in ComfyUI/models/sams/A SEGM detector model : like this one in ComfyUI/models/ultralytics/segm/
My own model merge and refiner setup
This is my own design. I worked through many iterations to land here. This is a heavy load workflow. Meaning, it will tap out a 12gb GPU. 10-30 second single image generations complete. I will not be updating this after I upload it. I am not taking requests for others either. This is just for my own tinkering and experimentation. Notes in the workflow can be found in Brown Groups. They are pretty informal. I assume, just by looking it over, most will be able to figure out how it works.Any use of this that is less than legal or problematic is on you. I am not responsible for misuse.
Ultra photo-realistic workflow | refiner, upscaling, FaceDetailer (+index elements), inpainting, controlnet, Civitai Metadata (NOOBAI-XL/ILLUSTRIOUS/SDXL) ComfyUi
STEPS BASE (after this value the refiner kicks in) should never exceed the value of TOTAL STEPS unless you want to use only the first KSampler.I like 10:14, increase total steps for realism, these are set for models in suggested resources, Uncanny valley and Fabled Illusion XXXL - (N/SFW), I really recommend downloading them, both are the best I found for their respective roles, feel free to comment your best values!The main appeal of this workflow is using illustrious/noobai-xl model to setup complex poses or characters and refine it with an sdxl photorealistic checkpoint.Version 5.5 has additional 1080p version file.I don't make new versions for small quality of life updates, download the file for the latest version.
SDXL Omni IMG Gen
A unified, modular SDXL/Pony/Illustrious powerhouse designed to consolidate professional-grade tools into a single, cohesive environment. This workflow eliminates "tab fatigue" by integrating CFG/Noise engineering, multi-LoRAs management, ControlNet, IPAdapter, InstantID, advanced Upscaling, and Detailers into ONE interface.Designed for speed and high-precision control, this workflow is essentially a "power workbench" overflowing with switches, sliders, and logic gates. It is undeniably complex, complex enough to make a beginner’s brain do a backflip, but for the pros and technical geeks - it’s a total playground. Think of it as the cockpit of a fighter jet: intimidating at first, but incredibly powerful once you know what the buttons do.Don’t panic! I’ve left a breadcrumb trail of notes and mini-tutorials in every single section. Please, I’m begging you read them. If you skip the manual and things get weird, don't say I didn't warn you! Devour the notes before you hit that "Queue Prompt" button; your VRAM (and your sanity) will thank you.Core Philosophy: Process in clear layers; fully configure the workflow in Row 1, generate and review a draft in Row 2, then produce the final high-quality render in Row 3. The completed image returns to Row 2 for side-by-side comparison before saving.The workflow is intentionally organized into three horizontal zones for logical progression:Row 1 (Top) – Setup & ConfigurationRow 2 (Middle) – Draft Generation & ReviewRow 3 (Bottom) – Final Processing & Output🛠️ System Requirements & HardwareGPU: NVIDIA (30xx/40xx Series).VRAM: 12GB Minimum / 24GB Recommended.System RAM: 32GB Minimum / 64GB+ Recommended.🎯 Software EnvironmentComfyUI Version: 0.15.0+ComfyUI Frontend: 1.39.16+💻 Core Software EnvironmentPython 3.12.x: (Highly recommended). While versions as low as 3.10, 3.11 may work, 3.12 provides the best compatibility for recent dependency updates). This version provides native support for nested f-strings and delivers improved execution speed.PyTorch: 2.10.0+cu130 - (Minimum 2.6.0+cu124) specialized build specifically engineered for full compatibility with CUDA 13.0.TorchVision: 0.25.0+cu130TorchAudio: 2.10.0+cu130📦 Essential Python DependenciesThese libraries are crucial for the stability of your Custom Nodes. Versions are strictly locked to prevent conflict.Core: numpy==1.26.4, setuptools>=70.0.0, pydantic>=2.10.0.Imaging: pillow==10.x, opencv-python, blendmodes==2025, glitch_this, rembg.Identity & Face: insightface==0.7.3, onnxruntime-gpu.Miscellaneous: tensorflow⚡ High-Performance AttentionPlease read this article for more information: https://civitai.com/articles/21516/install-sageattention-2-notes.SageAttention: Version 2.2.0 (Mandatory optimization).Flash Attention 2: (Fallback/Legacy support via Torch SDPA).Flash Attention 3: Version 3.0.0+cu130 (Optimized for RTX 40 series).Triton: 3.+ Windows-compatible build for FP8/NF4 dequantization.🧩 Required Custom Node PacksCore Infrastructure: rgthree-comfy, ComfyUI-Manager, ComfyUI_essentials, Use Everywhere (UE Nodes), ComfyUI-LogicUtils, ComfyMath, ComfyUI-JNodes, ComfyLiterals.CFG & Sampling: efficiency-nodes-comfyui, sd-perturbed-attention, Skimmed_CFG, RES4LYF, pre_cfg_comfy_nodes_for_ComfyUI, ComfyUI-Detail-Daemon.Identity & Perception: ComfyUI_IPAdapter_plus, comfyui_controlnet_aux, ComfyUI_ZenID, ComfyUI-WD14-Tagger, ComfyUI-AdvancedLivePortrait, Extra Models for ComfyUI.Upscaling & Motion: ComfyUI Impact Pack, ComfyUI Impact Subpack, UltimateSDUpscale, ComfyUI-SeedVR2_VideoUpscaler.Creative Suite: JPS Custom Nodes, Comfyroll Studio, ComfyUI-KJNodes, ComfyUI_Primere_Nodes, ComfyUI-JakeUpgrade, ComfyUI-pause, ComfyUI-mxToolkit.VFX & Realism: ComfyUI-Optical-Realism, CRT-Nodes, Virtuoso Nodes, ComfyUI_SKBundle, ComfyUI-FBCNN, comfyui_fill-nodes.📥 Model RequirementSDXL Checkpoint: Any SDXL/Pony/IllustriousSDXL VAE: Baked VAE (or any SDXL VAE)Upscale/ESRGAN: Any 4x upscale models.ControlNet: xinsir/controlnet-tile-sdxl-1.0; ttplanet_sdxl_controlnet_tile; diffusion_pytorch_model-instantID; diffusion_pytorch_model_union_promax (or you can use separated models for every preprocesses)IPAdapter: ip-adapter_sdxl_vit-h or ip-adapter-plus_sdxl_vit-h; ip-adapter-faceid-plusv2_sdxl ; ip-adapter-faceid-portrait_sdxl; ip-adapter-faceid-portrait_sdxl_unnormCLIP-Vision: CLIP-ViT-H-14 or CLIP-ViT-bigG-14Instant-ID: ip-adapter_instant_id_sdxlDetailers: Any detailer you prefer, I use person (#1), face (#2), eyes (#3), hair (#4), hands (#5), skin (#5), fashion (non-cond #1), furry (non-cond #2)Others: depth_anything_v2_vitl (depth map); ema_vae_fp16 (SeedVR2); seedvr2_ema_7b_fp8_e4m3fn_mixed_block/seedvr2_ema_3b_fp8_e4m3fn (SeedVR2); clipseg-rd64-refined-fp16 (noise map)🚩 Optimized Execution Flags (Optional)Use these flags in your .bat file to leverage the full power of your GPU--fast fp16_accumulation--dont-upcast-attention--use-sage-attention--disable-xformers (because xformers only support python <= 3.10.11)--preview-method auto--highvram/--normalvram/--lowvramIf you love my works, please review them with images and consider donating me some yellow buzz. This would be my great motivation to release more future works.
Regional Prompting with Regional LoRAs support
About v1.1ControlNet integration within regions can be tricky because ControlNet maps the reference image to the entire output image space, not to the mask space. This means that if the reference image contains a character in the center, the character will also be generated in the center of the output image, regardless of the mask.To address this, you can use one of the following approaches:Switch set_cond_area in the Regional Subgraph from the default setting to mask bounds (note: this may negatively affect interactions between regions),Or use the newly added "Pad image inside the mask" subgraph, which repositions and pads the reference image so it fits within the mask area.Also was added "Person Mask" subgraph, which allows extraction of a person mask from the input image for more precise region mapping.This subgraph uses a YOLO segmentation model, which should be placed in:/models/ultralytics/segm(you can download model here)Known Issues with the WorkflowFull mask coverage is requiredMasks must collectively cover the entire image. The combined masks from all regions should span the full image area. If any part of the image is not covered by at least one mask, that area will not be processed correctly during regional sampling and may appear empty, flat, or incorrectly rendered. If it is not possible to cover the entire image with masks, you can merge all masks into one, invert it, and use the result as an additional (background) region.Gradient (soft) masks are not supportedThe workflow does not work correctly with gradient (soft) masks. If a mask contains values below 100% (partial transparency), the sampler cannot properly denoise those areas. This results in visible noise, artifacts, or unstable textures. In some cases, you may reduce noise from non-binary masks by adding an additional KSampler with low CFG, step count, and denoise values—but this is not guaranteed to fully fix the issue.The workflow in theory should work with any base model regardless of architecture (for now it has been tested on SDXL models, Chroma, Flux, and ZImage).The workflow uses the recent ComfyUI subgraph feature, which allows extending the setup from 2 regions to any number of regions by simply duplicating the Regional Subgraph (and corresponding Regional FaceDetailer if needed).Regional Subgraph structureEach Regional Subgraph represents one independent region of the image and contains everything required to define a single character or element.The inputs of the Regional Subgraph are:region maskDefines where this region is applied in the image. This can come from any mask generation method (rect masks, segmentation, depth, manual paint, etc.).regional positive conditioningThe prompt that describes the character or object inside this region.regional negative conditioningRequired, but can be either global negative prompt or region-specific.model (branched)This is a key part of the workflow. Instead of using the base model directly, the model is first branched and modified (for example by applying LoRAs) before being passed into the Regional Subgraph.Model branching and region-specific LoRAsEach region has its own model branch. This is done by taking the base model and applying region-specific LoRAs before feeding it into the Regional Subgraph.Typical setup:Base model with global LoRAs → shared for the whole image and passed directly into Regional SampleFor each region:duplicate the model pathapply one or more LoRAs specific to that regionpass the modified model into the Regional Subgraph model inputThis allows each region to have completely different:charactersstylesLoRAsvisual identityImportantly, this avoids the common problem where LoRAs are applied globally and interfere with each other.Extending the workflowTo add more regions:Copy an existing Regional SubgraphConnect:a new maska new prompta new model branch (with its own LoRAs)Add the output to:Combine ConditioningCombineRegionalPromptsThis makes the system fully scalable without restructuring the graph.CustomizationEverything except the following can be replaced depending on user preference or task:Regional Subgraphs and Regional Sampler (structure must remain for modularity)Combine ConditioningCombineRegionalPromptsUsers are free to replace:Mask generation (rect masks, segmentation, ControlNet, etc.)Prompt encoding pipelineLoRA setupFaceDetailer or any post-processingThis allows adapting the workflow to:character scenesobject compositionstyle mixinginpainting pipelinesanimation setupsImportant note: Base steps vs Regional samplingOne of the most important parameters in this workflow is base_only_steps, which controls how the generation process is split between:Base sampling stageRegional sampling stageHow it worksThe total number of steps is defined by the sampler (for example 30 or 50 steps).These steps are divided into two phases:Base sampling (base_only_steps)Only the global prompt is appliedNo regional prompts or masks are used yetThis stage defines:compositionposecamera anglegeneral structure of the imageRegional sampling (remaining steps)Regional prompts and masks are appliedEach region modifies its assigned part of the imageThis stage refines:character identitylocal detailsLoRA-specific featuresWhy this mattersThe base stage essentially “locks in” the structure of the image.If too many steps are spent in base sampling:the image becomes stable and coherentbut regions have less ability to change itLoRAs may appear weak or not recognizableIf too few steps are spent in base sampling:regions have strong influencecharacters become more accuratebut overall composition may become unstable or inconsistentRecommended defaultA good starting point is:base_only_steps = 50% of total stepsThis provides a balance between:stable compositioneffective regional controlWhen to increase base_only_stepsIncrease base steps if:composition is broken or inconsistentcharacters are not aligned properlyperspective or layout is unstableThis gives the base stage more time to establish a solid structure.When to decrease base_only_stepsDecrease base steps if:region-specific LoRAs are weakcharacters are not recognizableregional prompts have little effectThis allows the regional stage to have more influence over the final image. However prefer increasing total steps first before decreasing base_only_stepsPractical intuitionBase stage = “what the image looks like”Regional stage = “who/what is inside each part”Balancing these two is key to getting both:strong compositionstrong character identity
Illustrious ADetailer workflow for ComfyUI
This is the battle-tested NoobAI/Illustrious workflow I use to generate characters. It imitates the ADetailer feature to upscale the generated image and do some fix on character faces and hands.Warning: since the developer of RvTools does not support Nodes 2.0 they should be disabled in ComfyUI general settings.Required ComfyUI custom nodes for the workflow:ComfyUI-Impact-Packrgthree-comfyComfyUI-Easy-UseComfyUI_essentialsComfyUI-IcyHiderComfyUI-RvTools_v2ComfyUI-Lora-Manager
Foton's Workflow with ControlNet upscale
Simple workflow with ControlNet tiled upscaleThis is my main workflow for SDXL/Pony/Illustrious that uses ControlNet tile upscaling instead of Ultimate SD. Feel free to DM me if you have any questions about the workflow.v1.2 updateAdded CFG Skimming, after more testing I think it's an interesting tool. Can be easily disabled in generation settings options (enable_skimming). Change CFG to recommended checkpoint values when you disable it.Saved images should now be compatible with Civitai. I was getting tired of filling some fields manually. Requires manual installation of Save Images With Metadata Universal ExtensionCombined generation settings into a single subgraphFor MetaData saver to work properly, run Rules Scanner Subgraph once after installation or when you change nodes in ComfyUI (I don't make the rules)By default set tile_batch_size to 1 in HiRes subgraph to safely run on <12GB VRAM. If you have more than 12GB, set it to 2 (can be even higher with 24GB, but I won't recommend it).Removed redundant switch for Adetailer. v1.1 updateReplaced Lora Loaders to non-Clip onesReplaced KSampler to default ComfyUI nodeCleaned up wiringPromoted additonal options in HiRes subgraph for easier accessInitial upscale model selector (i.e. Remacri/UltraSharp/ESRGAN)ControlNet tile checkpoint selectortile batch size (higher values increase speed at the cost of VRAM)blend factor (color correction)Also added some options in the face detailer subgraphMoved nodes closer together and grouped them for easier bypassFeaturesSlower than Ultimate SD upscaler but produces better quality in my opinionAllows for higher denoise without adding weird detailsEasily switch HiRes and FaceDetailer subgraphs on and offSave upscaled output as PNG and JPG, I usually post JPG images for faster loading. PNG versions will include workflow in a metadata.Can be VRAM hungry, only tested it on my 12GB cardUses additional ControlNet tile model:Realistic version, but it works well even for stylized outputs - good results with fabric, fur and skin texture, but tends to oversharpen hair (can play with strength settings to fix this): https://huggingface.co/TTPlanet/TTPLanet_SDXL_Controlnet_Tile_RealisticAlternative model for anime and stylized models - sharper and cleaner details: https://huggingface.co/xinsir/controlnet-union-sdxl-1.0Put these models into ComfyUI/models/controlnet folderKSampler can be replaced with a ComfyUI one, I'm using Inspire's node for different noise generation options.Control upscale size and denoise in the HiRes subgraphFor higher upscale values (3-4x) use lower denoise (0.2 or even lower).Custom Adetailer promptIn Use custom prompt node set value to true and type positive/negative prompts which will be used for faces. Userful when some Loras overpower Adetailer and change face color too much, or if you just want to adjust face expression. With additional nodes you can set it up to use different Loras during this step. Increase feather value if you get visible edges around faces.Low VRAMTo save VRAM inside a HiRes Upscale subgraph try this:In Tiled Diffustion node Set batch size to 1. This will reduce VRAM load considerably at the cost of speed.Set VAE Decode tile size to 512.If it's still VRAM hungry, try setting tile width/height in Tiled Diffusion and VAE encode to 512x512. Might want to change overlap as well, but I haven't tested this.
Illustrious/Pony/SDXL essentials
Note: I spend a lot of time on these to make sure they are top notch. If you have ANY issues, I'm very responsive. hit me up and I'll resolve them for you.A complete user and troubleshooting guide inclusing a list of common error messages can be found here https://civitai.com/articles/24678/how-to-use-my-workflows-a-comprehensive-breakdown也有中文说明Instagram: https://www.instagram.com/synth.studio.models/Buy me a☕ https://ko-fi.com/lonecatoneThis represents hundreds of hours of work. If you enjoy it, please 👍like, 💬 comment , and feel free to ⚡tip 😉This is a pruned version of my main workflows. I designed it for beginners so that they could understand the basics. The full version is here: https://civitai.com/models/2189190Sage Attention: Speeds up productionUltimate Upscaler for prescaling and hi res fixDetailing Suite for face, hands, eyesSeed VR2 Upscaler to get 4k on low VRAMEssential Post Production SuiteJPEG ScrubberFull tutorial available.也有中文说明
Cappy's SwarmUI Presets
THESE ARE NOT COMFYUI WORKFLOWS!!! THESE ARE MADE FOR SWARMUI!!!!!I made these presets because I want to speedrun my shit and not have to set up everything each time... I'm sharing these presets in case anyone finds them useful.Make sure to read the " About Version " section
Illustrious XL & Multi-Model Refiner: From Creativity to Specialized Style
This ComfyUI workflow is designed for creators who want to leverage the massive creative potential and prompt adherence of Illustrious XL, while injecting the specific artistic textures of more specialized SDXL checkpoints.The core philosophy is "Checkpoint Morphing": using a base model for "The Soul" (composition/concept) and a specialized model for "The Skin" (textures/line-work).Key Features:Dual-Pass Generation: Uses a high-denoise pass on Illustrious XL to establish the "creative bones" of the image.Latent Refinement: A second KSampler pass with a lower denoise (optimized between 0.45 - 0.6) to "re-skin" the image using a second specialized checkpoint.Artistic Preservation: Specifically tuned to maintain the silhouette and character likeness while transforming digital renders into traditional media styles like watercolor, ink-wash, or sketch.Universal Compatibility: Works with any SDXL-based checkpoints, allowing for endless stylistic combinations.How to Use:Checkpoint 1: Load Illustrious XL (or your favorite creative base).Checkpoint 2: Load a Specialized Model (e.g., Watercolor, Sketch, or Cinematic styles).Prompting: Use your main prompt in the first stage. You can add style-specific trigger words in the second stage for better results.Denoise Control: * 0.3 - 0.4: Subtle texture enhancement.0.5 - 0.6: (Recommended) Balanced stylistic transformation.0.7+: Major compositional shifts (high creativity, low fidelity to the original).Requirements:ComfyUI-Custom-Scripts (for better node management).Impact Pack (optional, but recommended for face detailing).🚀 Pro-Tip for Users:If you find the composition shifting too much between passes, try lowering the denoise of the second KSampler. If the style isn't "bleeding" through enough, increase it slightly or add a style-specific LoRA to the second model.
Simple Merge model, 4x adetailer
This is a simple adetailer workflow with 4x face detailer nodes and model merger. I feel this is a good starting point for noobs such as myself. should be good with Illustrious, Noobai, Pony, atleastCustom nodes are Comfyui Impact pack and Comfyui Impact subpack
Celestial thighs workflow
Video of me using the workflow:Credit to Lordgreeman for the original workflow, this is a tweaked version of his.
Illustrious Workflow
A Illustrious workflow that I use to handle kewords, loras, upscale and lora testing
Anima_to_SDXL/IL/Pony_with_face_detailer
🧸Anima → SDXL Conversion Workflow Keep Pose, Upgrade QualityGenerate with Anima, finish with SDXL / IL, keep the pose, keep the composition, upgrade the quality.This workflow lets you use Anima for structure and SDXL for rendering without losing character layout.Perfect for people who want:Anima pose stabilitySDXL detail qualityLoRA compatibilityCharacter consistencyEasy multi-model workflowNo manual latent editing needed.Just load → run → done.Includes FaceDetailer and automatic save at every stage.✨ What this workflow doesGenerate base image with AnimaConvert latent to SDXLRe-render with SDXL / IL modelKeep pose / composition / characterApply FaceDetailer automaticallySave before / after imagesOutputsAnima beforeAnima afterSDXL beforeSDXL after🧠 How it worksGenerate with AnimaDecode → Encode using SDXL VAERun SDXL sampler with same promptFaceDetailer refinementSave resultsStructure stays the sameRendering gets upgraded🚀 Why you want thisFix low quality Anima imagesKeep pose when switching modelsTest LoRA on different checkpointsConvert Anima to IL / IllustriousMulti-model workflow made easyStable character generationGreat for animation / dataset / LoRA testingThis workflow is especially useful if you useAnimaSDXLPonyIL / IllustriousSmudgeGardencustom LoRARequired ModelsAnimaAnima modelhttps://huggingface.co/circlestone-labs/Anima/tree/main/split_files/diffusion_modelsText encoderhttps://huggingface.co/circlestone-labs/Anima/tree/main/split_files/text_encodersVAEhttps://huggingface.co/circlestone-labs/Anima/tree/main/split_files/vaeLoRA (example: fully_plush_doll_with_anime_eye_AN_v1)https://civitai.com/models/825194?modelVersionId=2732496SDXL / ILModel (example: SmudgeGarden_IL_v1.527)https://civitai.com/models/2393702/smudgegardenilVAE (example: liquid9745vae)https://civitai.com/models/85106/liquid111vae-and-liquid9745vaeLoRA (example: fully_plush_doll_with_anime_eye_AN_v1)https://civitai.com/models/825194/fullyplushdollwithanimeeyesRequired ExtensionsComfyUIhttps://github.com/comfyanonymous/ComfyUIComfyUI Managerhttps://github.com/ltdrdata/ComfyUI-ManagerImpact Packhttps://github.com/ltdrdata/ComfyUI-Impact-PackImpact Subpackhttps://github.com/ltdrdata/ComfyUI-Impact-SubpackUltralyticshttps://github.com/ultralytics/ultralytics
ComfyUI - simple and basic Image Workflow | text2image | lora | control net | wildcards support 🔖
ComfyUI – Simple & Basic Image Workflow 🎨An easy-to-use workflow for image generation:• text2image• LoRA support• ControlNet• wildcards support 📜Perfect for beginners.📚📃Notes⚠️ If wildcards are NOT enabled, write your positive and negative prompts in the bottom fields, not the top ones.✍🏻 📃Safe rule✅ With wildcards:Positive → TOP || Negative → also TOP✅ Without wildcards:Positive → BOTTOM || Negative → BOTTOM🔛 Example for Ksmaplers 🔛This setup may produce images with a slight anime-style look, similar to the showcase image.⚠️ Note: If anatomical errors occur, tweak the HiresFix percentage (usually lowering it helps).⚜️Lora's location-----ComfyUI_windows_portable\ComfyUI\models\loras⚜️Upscale location----ComfyUI_windows_portable\ComfyUI\models\upscale_modelshttps://civitai.com/models/147759/remacri⚜️ControlNET location-----ComfyUI_windows_portable\ComfyUI\models\controlnethttps://civitai.com/models/136070/controlnetxl-cnxl⚜️Wildcards location-----ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-impact-packhttps://civitai.com/models/2111628?modelVersionId=2388856
ComfyUI workflow with Regional Prompt for couple focus V3
Regional prompt workflow for XL ModelsThis is a combination of some workflow I'v used in the past or recently.Most of the part from the regional prompt was originally on this workflowHuge thx to @zml_w for all the testing and the helpI'v just found that Levpat might be banned or deleted account Upscale Method for V2 by @LevPat https://civitai.com/models/1994653/custom-upscale-flowUSER GUIDEhttps://civitai.com/articles/18257For V3 Merged WorkflowMerged the two method into one workflow, you can now toggle between Dense Diffusion and Attention coupleChanged the detailer method, I'v kept Couple Face Detail group (for some image it's ok)Now I'm mostly using a method from MGHerderFound on this workflowAdded some settings to the Face detailer, I'v linked all the merged text prompt to the positive on the first two SEGS for each two characters, you can swicth between character (I couldn't find a good way to had wildcard or overwrite prompt on the SEGS detailer)Added SeedVr2 upscale you will need to download the model and VAE for it to workAdded a more simple and lighter latent UpscaleI'v kept the Advanced latent upscale (was called Crazy Upscale before) linked all the text positive and negative prompt to it, converted to subgraph all the noodlesKept the USDU upscale in some use it can be good tooBasically you have now 4 upscale method to choose with, you can delete the Upscale group you don't want it will not break the workflow.You will need all the the custom nodes below that was used for V2 and all the followinghttps://github.com/numz/ComfyUI-SeedVR2_VideoUpscaler (If you want to use it)https://github.com/mcmonkeyprojects/sd-dynamic-thresholdinghttps://github.com/chrisgoringe/cg-use-everywherehttps://github.com/BlenderNeko/ComfyUI_ADV_CLIP_embhttps://github.com/LAOGOU-666/Comfyui-Memory_Cleanuphttps://github.com/Extraltodeus/Skimmed_CFGhttps://github.com/mirabarukaso/ComfyUI_Mirahttps://github.com/ltdrdata/was-node-suite-comfyuiFOR V2 Dense Diffusion and Attention CoupleI'v made two separate workflow with different Regional prompt method, both of them are in the archive if you want to try.Dense DiffusionSo the error "The size of tensor a (924) must match the size of tensor b (308) at non-singleton dimension 3"Is still here with Dense Diffusion BUT now I know 100% why it happens, so I will detail everythings how to avoid it.First you can put almost everything in every positive prompt, a lot of detail or action for left char and almost nothing for 2nd character will not trigger this error.The only way this error will pop is Positive VS Negative Prompt.So why it happens then, almost exclusivly with Embedings, because of Tokens count and the way Dense Diffusion work, the size of Tokens need to be almost the same, and when you put emebeddings, like Lazyhand, Lazyneg, etc will use a lot of Tokens (more exactly tokens Chunks)So how to avoid it, you can either not use emebeddings at all, I'v added note on the workflow with good (I think) quality prompt.You can still use Embeddings but you wiil need to balance them between positive and Negative. So for example if you put Lazyhand + Lazyneg in Negative, you will need to use at least one or more embeddings in the positive quality to get the same size of Tokens chunks.So because of that I'v added a custom nodes made by me (and GPT) to Normalize the negative quality, you can either turn it off or on, I'v put pics for comparaison in the archive.If you'r not happy or don't like this method I'v made another Workflow with a different nodes called Attention couple.Attention CoupleWith Attention Couple this error will never pop, then why not using it and getting rid of Dense Diffusion ?Attention couple work great and can have pretty good result, but in my testing I saw two major flaw with it.First the Lora did not get applied the same way with it, Dense Diffusion seems to have a better blend for the model and loras.Second the interaction seems more difficult too, I did not use it very extensively but the few image I'v made seems to have less prompt adherence.Here is every nodes you will need for bothFor Dense Diffusionhttps://github.com/Fannovel16/comfyui_controlnet_auxhttps://github.com/ltdrdata/ComfyUI-Impact-Packhttps://github.com/rgthree/rgthree-comfyhttps://github.com/yolain/ComfyUI-Easy-Usehttps://github.com/kijai/ComfyUI-KJNodeshttps://github.com/ssitu/ComfyUI_UltimateSDUpscalehttps://github.com/cubiq/ComfyUI_essentialshttps://github.com/ClownsharkBatwing/RES4LYFhttps://github.com/giriss/comfy-image-saverhttps://github.com/shiimizu/ComfyUI_smZNodeshttps://github.com/ltdrdata/ComfyUI-Impact-Subpackhttps://github.com/huchenlei/ComfyUI_densediffusionhttps://github.com/edelvarden/comfyui_image_metadata_extensionhttps://github.com/rcsaquino/comfyui-custom-nodeshttps://github.com/zml-w/ComfyUI-ZML-Image(https://github.com/zml-w/ZZZ_ZML_English_Patch)A small custom node is in the archive for the Negative Normalization, just need to put it in custom nodes folder. No requirements needed.For Attention CoupleEverythings above +https://github.com/BlenderNeko/ComfyUI_ADV_CLIP_embhttps://github.com/laksjdjf/cgem156-ComfyUIIf you have trouble with FAILED IMPORT for cgem156 you can try this fixhttps://github.com/laksjdjf/cgem156-ComfyUI/issues/17#issuecomment-2918745574Worked for me with ComfyUI desktop AppOlder VersionI'v added a chain sampler because I love this method, the render time and the result it gave.This is quiet an advanced workflow so you will need a lot of custom nodes, some may not work with ComfyUI manager, I will link them here if you want to install manually.Make sure you activate your venv environment for the requirements.txt for custom nodes.There is a simplified version in the archive if you want to try without all the nodes for HiresFix, Face detailer, upscale and color match.I'm not responsible if you break your ComfyUI install, some of the nodes may use different Pytorch version or requirements you already use.https://github.com/ltdrdata/ComfyUI-Impact-Packhttps://github.com/pythongosssss/ComfyUI-Custom-Scriptshttps://github.com/rgthree/rgthree-comfyhttps://github.com/yolain/ComfyUI-Easy-Usehttps://github.com/shadowcz007/comfyui-mixlab-nodeshttps://github.com/jags111/efficiency-nodes-comfyuihttps://github.com/ssitu/ComfyUI_UltimateSDUpscalehttps://github.com/Suzie1/ComfyUI_Comfyroll_CustomNodeshttps://github.com/cubiq/ComfyUI_essentialshttps://github.com/ClownsharkBatwing/RES4LYFhttps://github.com/ltdrdata/ComfyUI-Impact-Subpackhttps://github.com/huchenlei/ComfyUI_densediffusionhttps://github.com/edelvarden/comfyui_image_metadata_extensionhttps://github.com/Miosp/ComfyUI-FBCNNhttps://github.com/rcsaquino/comfyui-custom-nodeshttps://github.com/weilin9999/WeiLin-ComfyUI-prompt-all-in-oneDeprecated version use this one instead, I'v fixed the workflow with the new onehttps://github.com/weilin9999/WeiLin-Comfyui-ToolsFor V1 Multimaskhttps://github.com/chflame163/ComfyUI_LayerStylehttps://github.com/kijai/ComfyUI-KJNodeshttps://github.com/chrisgoringe/cg-use-everywherehttps://github.com/giriss/comfy-image-saver
LewdSpirit's txt2img Workflow
Decided to create my own workflow, trying my best to keep it simple and compact. Let me know what you think, it works flawlessly for me so far. It includes face detailer that can be toggled either on or off.EDIT 3/15Uploaded a clean, neater version of the workflow.
illustrious TxT2ImG with Controlnet, Detailer Deamon, Detailers and IPAdapters
🗒️ My NEW Illustrious Workflow for ComfyUIHi, so I thought I’d share the Illustrious Workflow i've been using for a while now.This is an evolution of my last workflow which you can find here .Main Points of FunctionMultiple Detailers: Person, Face, Eyes, Hands, Breast, and Nipples. Each individually toggleable using the bypassers, so you can fine-tune exactly what gets refined.Selectable ControlNet System: Turn on/off entire ControlNet groups or activate specific nets when you need targeted structure or pose control.IPAdapter Integration: Supports multiple reference images at once, each with adjustable weight, for style, lighting, and character consistency.Group LoRA Loaders: Pre-load entire packs of LoRAs with weights, flip them on/off instantlyDetailer Deamon: Sharpen textures, enhance fine features, or smooth things out, without touching your samplers or prompts.Two-KSampler Workflow: A flexible dual-sampler setup that gives you extra shaping control without complicating your workflow.Full Metadata Saving: Automatically stores checkpoint, prompts, settings, and any LoRAs used in the prompt, ideal for sharing and CivitAI uploads.Auto Saving: Add %date to your save path and ComfyUI will neatly organise your outputs into daily folders.Beginner-Friendly Bypass Architecture: Start simple, then switch on advanced features one at a time — ideal for learning without overwhelm.SAM + Ultralytics Segmentation Support: High-quality segmentation models ensure your detailers see what they’re supposed to see — faces, hands, bodies, and more.DisclaimerI’m no expert, in fact, i'm a bit of a novice. This workflow evolved from my last workflow, which in turn came from looking at other great ones by DigitalPastel and Mittoshuras, then I linked things together until they worked.Workflows I built from:https://civitai.com/models/1669611/dynamic-multi-character-workflow-regional-prompts-and-wildcards-inside-comfyuihttps://civitai.com/models/1598938/smooth-workflow-txt2img⚠️ That means if you have trouble running this workflow, I'm afraid I probably won’t be able to support you as much as I’d like to, or even at all really.I’ll write a short a step-by-step guide below, but over the next week or two, I intend to write a full article detailing the workflow in full, but, in the mean-time, if you are running into issues, ChatGPT is your friend (that’s what I did) copy and paste your error into that, and stick with it.Alternatively, if you look at this workflow and think “what the hell is this” I seriously suggest DigitalPastels Smooth-workflow, It was an absolutely great start for me.⚡ Quick StartQuickly start creating images in the most beginner friendly way:1. Load the workflow into ComfyUI.Hide links, this workflow is a birds-nest of noodles.2. Leave everything OFF except the Image Creation section.This keeps things clean and beginner-friendly, or just to test the workflow on your rig. Use the main Bypasser to make sure everything is off except for Txt2Img.3. Set your Image Saver PathIn the imager saver node, set your image name, path and format to be saved.Currently, images will be saved in a folder named Text2Img and then in a sub-folder named as per the date the image was created, for example, Text2Img/2025-12-01/filename.png4. Select your checkpoint.5.Load a VAE (Only If You Need To)Load a VAE only if you actually need one. Do this by connecting the Load VAE node to the VAE Anything Everywhere node. (As default, this node will be connected the the Checkpoint loader)6. Set an image size using the 2D slider.7. Write your positive prompt.Use the Green Positive Prompt node. You can either keep it simple for a first test, e.g.“a detailed portrait of a woman, soft lighting, photorealistic”or if you'd like to go a little more advancedThe Positive Node is divided into 2 areas.Top box = your positive prompt.Use "Select to add LoRA" to insert LoRAs.Format examples:embedding:SmoothNoob_Quality<lora:Smooth_Booster:0.5>6. Add a basic negative prompt.Use the Red Negative Prompt in the same way as positive including embeddings, or keep it simple to start, e.g.“blurry, distorted anatomy, extra limbs, bad hands”7. Press Queue Prompt.Congrats, your first output is done.8. Want more control?Turn new sections on one at a time:DetailersControlNetsIPAdapterLoRA Loader GroupsThis workflow is built so you can scale up gradually without breaking anything, or should I say when something does break, it should be easy to roll back.Below is a breakdown of what each one does, and more importantly, when you should consider using them.⭐ Detailers – What They Do & When to Use ThemThis workflow includes six dedicated detailers designed to clean up and refine specific parts of your image. Each one can be toggled on/off using the bypassers, so you only activate the ones you actually need.Think of detailers as your “post-process artists,” stepping in to refine the areas that models often struggle with.👤 Person DetailerThe Person Detailer refines the whole character at once. It cleans up broad shapes, clothing edges, hair outlines, and major anatomical features before the more specialised detailers take over.When should you use it?Use the Person Detailer when:• The overall character looks mushy or soft• Clothing or body edges need sharpening• Hair looks blobby or lacks structure• The base render has good composition but weak definition• You want a clean foundation before face/hands/eyes refine furtherThis is a great “first pass” detailer — especially for portraits or full-body shots.🙂 Face DetailerThis one targets the entire face area: contours, nose, lips, jawline, expression, cheek details — the whole package.When should you use it?Use the Face Detailer when:• The face looks blurry or undefined• Expression clarity is important• You want better skin detail without harsh artefacts• You’re working with portrait-style shots• The base model tends to give inconsistent facesAlmost always worth turning on if the character is important.👁️ Eye DetailerEyes are the centrepiece of most images, and models often smudge or distort them. This detailer isolates and refines only the eyes and eyelids for maximum sharpness.When should you use it?Use the Eye Detailer when:• Eyes look soft, uneven, or mismatched• Iris details don’t pop• Eyelashes turn into fuzzy shadows• You want more expression or intensity in the gaze• You’re generating close-ups or portraitsThis detailer dramatically improves perceived quality with minimal processing.There is also a checkpoint loader here, I have found that mixing Checkpoints when it comes to eyes can produce some great, dramatic, results. ✋ Hand DetailerHands may be the final boss of AI image generation. This detailer helps keep them under control by refining the mask and sharpening finger shapes.When should you use it?Use the Hand Detailer when:• Fingers are smudgy, fused, or unclear• You’re working with poses where hands are visible or prominent• Jewellery, tattoos, or props are involved• You want more realistic hand shading and structure• Close-ups include hands near the faceIf hands appear in frame — this detailer becomes your best friend.⭐ Breast DetailerRefines shapes, shading, and contours in the chest area. Helps correct deformities or modelling quirks, especially at higher resolutions or stylised character designs.When should you use it?Use the Breast Detailer when:• You’re working with NSFW, glamour, or revealing outfits• Clothing fabric bunches incorrectly• The chest area has stretching, warping, or odd shadows• You want smoother curvature and more natural highlightsOptional, but extremely helpful for figure-focused compositions.🔸 Nipple DetailerThis detailer is highly dependent on the segmentation model and may require experimentation. It refines only the nipple region and can correct common texture or placement issues, but results vary between models.When should you use it?Use the Nipple Detailer when:• You’re generating NSFW content• The model’s default nipple rendering is inconsistent• You need sharper detail or more realistic texture• You’re using a checkpoint that struggles in this specific areaIf your model already renders nipples well, you can leave this off. Otherwise — useful to experiment with.🎛️ General Advice for Detailers• You don’t have to use all detailers at once — activate only the ones relevant to your image• For portraits: Face + Eyes gives the biggest boost• For full-body shots: Person + Hands is often enough• For NSFW: Breast + Nipple + Person creates the strongest refinements• If the image looks over-processed, turn one or two offDetailers shine brightest when used strategically — not constantly.⭐ Detailer Deamon – The Secret Sauce for Smarter DetailThe Detailer Deamon (from Impact Pack) enhances image refinement by adjusting the sigma curve, the noise schedule your sampler follows during denoising. By shaping this curve, the Deamon gives you more control over when detail appears during the render.This allows you to:• Add structural detail earlier in the denoising process• Add fine, crisp detail later for a cleaner finish• Reduce oversharpening or muddy textures• Improve overall texture handling across SDXL, SD1.5, FLUX, and moreThink of it as a detail timing controller: instead of relying solely on sampler settings, the Deamon fine-tunes the flow of detail so your images look smoother, cleaner, and more intentional.When should you use this?Turn on the Detailer Deamon when:• Your images feel too soft or lack definition• Your detailers are producing results that look harsh or oversharpened• You want smoother, more consistent refinement without touching complex sampler settings• You're working at high resolutions and want better control over fine textures• You want more “polish” without increasing steps or using external upscalersIt’s optional — but once you get used to the cleaner detail flow, it’s hard to go back.How to install it:Download Impact Pack:https://github.com/ltdrdata/ComfyUI-Impact-PackPlace it into:ComfyUI/custom_nodes/ Restart ComfyUI — and the Detailer Deamon is ready to refine your renders.⭐ ControlNets – Installing and Understanding the Selectable SystemControlNets are your workflow’s “structure helpers.” They let you guide the model using external information like pose, edges, depth, lineart, or scribbles. Instead of hoping the model interprets your prompt correctly, ControlNets give it a gentle nudge (or a firm shove) in the right direction.This workflow includes a Selectable ControlNet System, which keeps everything tidy and easy to manage:• Toggle entire ControlNet groups on or off• Enable only the individual nets you need• Prevent your graph from turning into a ControlNet spaghetti monsterWhere to download SDXL ControlNets:• https://huggingface.co/lllyasviel• https://huggingface.co/InstantX/• https://huggingface.co/thibaud• https://civitai.com — search “SDXL ControlNet”Where to put them:ComfyUI/models/controlnet/ Once installed, they’ll appear automatically in the dropdown menus inside each ControlNet Loader.How the selectable system worksMain blue box → activate or disable the entire ControlNet sectionYellow boxes→ turn individual ControlNets on/offEverything stays organised instead of stretching across half your screenStart simple and switch on additional nets only when your project needs them.⭐ When should you use ControlNets?Turn on a ControlNet when:• You want a specific pose and don’t want to fight the RNG• You have a reference sketch or scribble you want the model to follow• You need cleaner structure, straight lines, defined shapes, consistent silhouettes• Your images keep drifting away from the composition you intended• You want depth or edge information to guide lighting and shading• You're working with character consistency and want them to stay on-model• You want more control without writing enormous promptsThey’re optional, but, incredibly powerful once you get used to them.And with the selectable system, they’re easy to toggle on and off as needed.⭐ IPAdapter – What It Is & How to Install ItIPAdapter is your workflow’s reference interpreter. It allows you to feed an image into the pipeline and guide the model toward matching its style, composition, lighting, or character identity without overpowering your text prompt.Unlike ControlNets, which enforce structure, IPAdapter works more like a gentle influence:“Make it look a bit more like this.”This workflow includes multiple IPAdapter slots so you can blend several references at once, each with its own weight for fine control.Where to download IPAdapter (SDXL version):• https://huggingface.co/h94/IP-Adapter/tree/main/sdxlDownload the .safetensors files for SDXL.Where to download CLIP Vision models (required):• https://huggingface.co/laion/CLIP-ViT-H-14-laion2B-s32B-b79KWhere to put them:ComfyUI/models/ipadapter/ ComfyUI/models/clip_vision/ Once placed correctly, they’ll appear in the IPAdapter nodes automatically.⭐ When should you use IPAdapter?Turn on an IPAdapter when you want your image to follow visual cues from a reference rather than purely relying on text.It’s especially useful when:• You want to match a character’s face, hairstyle, or silhouette• You want consistent character identity across multiple images• You prefer a specific lighting style or colour palette• Your prompt alone struggles to achieve the look you want• You want to blend multiple influences (e.g., style from one image, lighting from another)• You’re working with cosplay, portraits, or anything requiring recognizable likenessIPAdapter is also great when you need guidance without strict ControlNet structure, it influences the overall feel of the image rather than forcing exact shapes.If you only need style or vibe matching, IPAdapter is often all you need.If you need pose or precision structure, pair it with a ControlNet.⭐ LoRA Loader Groups – The Fastest Way to Use Multiple LoRAsLoRA Loader Groups let you preload whole sets of LoRAs and toggle them on or off instantly, without cluttering your positive prompt. They’re ideal when you want to switch between styles, character identities, or thematic looks quickly and consistently.Instead of manually typing LoRA weights like <lora:thinghere:0.8> each time, these loaders allow you to:• Pre-load several LoRAs at once• Adjust their weights in one place• Turn the entire group on/off using bypassers• Keep your main prompt clean and readableEach group also includes a handy string box where you can store trigger words, tags, or character names associated with those LoRAs, making copy/paste into your prompt fast and painless.This system is especially helpful in complex workflows where prompt clutter becomes a real issue.⭐ When should you use LoRA Loader Groups?Use LoRA Loader Groups when:• You frequently switch between different styles (anime, realism, cinematic, painterly)• You have multiple character LoRAs and want to swap them quickly• You’re experimenting with theme bundles (e.g., “cyberpunk pack,” “fantasy armour pack”)• You want consistent LoRA weights across multiple renders• You want to avoid clogging your prompt with long LoRA syntax• You’re running batch generations where different LoRA configurations are requiredThey’re especially powerful in workflows like this one, where modular toggling is a core design principle.If you only need one LoRA occasionally, adding it directly in the positive prompt is fine.But if LoRAs are a regular part of your generation process, Loader Groups save time, reduce inconsistency, and keep everything organised.I'm really looking forward to seeing what people can create with this workflow. If you use it, and decide to post your creations to Civitai, please consider posting via the "Gallery > Add Post" feature, so I can see your creations.