ComfyUI Workflows
Browse real AI workflows from Flux, SDXL, Pony, ControlNet, video generation and commercial AI creators.
FLUX.2 Klein 9B - Reference-Based Face Swap Workflow
FLUX.2 Klein 9B - Reference-Based Face Swap Workflow⚡ Production-tested workflow.This workflow originated from an internal automation pipeline capable of processing hundreds of images automatically through the ComfyUI API on a single NVIDIA T4 GPU.A highly consistent face swap workflow for ComfyUI built around FLUX.2 Klein 9B, latent references, and a dedicated face swap LoRA.Unlike traditional face swap methods, this workflow uses both a Target Reference and a Source Face Reference to preserve composition while transferring identity.RequirementsBefore loading the workflow, make sure you have:FLUX.2 Klein 9BFLUX2 VAEQwen 3 8B FP8 Mixed Text EncoderFace Swap LoRALanPaint Custom NodesRequired ModelsFLUX.2 Klein 9BPlace inside:ComfyUI/models/unet/ Model:flux-2-klein-9b.safetensors FLUX2 VAEPlace inside:ComfyUI/models/vae/ Model:flux2-vae.safetensors Qwen 3 8B FP8 MixedPlace inside:ComfyUI/models/text_encoders/ Model:qwen_3_8b_fp8mixed.safetensors Face Swap LoRAPlace inside:ComfyUI/models/loras/ Model:bfs_head_v1_flux-klein_9b_step3500_rank128.safetensors Required Custom NodesLanPaintThis workflow uses the LanPaint sampler.Install it through:ComfyUI Manager → Install Missing Custom NodesWorkflow OverviewThis workflow uses two different references:Target ReferencePreserves:CompositionPoseFramingPerspectiveScene structureLighting contextSource Face ReferenceTransfers:Face identityFacial proportionsHair characteristicsEyesSkin detailsBy combining both references, the workflow achieves significantly better consistency than traditional face swap approaches.How To UseStep 1 - Load The Source FaceNode:Source Face ImageThis image provides the identity that will be transferred.Recommended:✅ Front-facing portraits✅ High-resolution images✅ Clear facial visibility✅ Good lightingAvoid:❌ Motion blur❌ Heavy occlusions❌ Sunglasses❌ Extreme anglesStep 2 - Load The Target ImageNode:Target ImageThis image provides:BackgroundBodyClothingPoseCompositionFramingPerspectiveThe workflow attempts to preserve these elements while replacing the identity.Recommended:✅ Portrait photos✅ Medium shots✅ Visible face✅ Clear lightingAvoid:❌ Tiny faces❌ Hidden faces❌ Extreme side profilesStep 3 - Queue PromptPress:Queue PromptThe workflow automatically:Reads target dimensionsCreates target latent referencesCreates source face latent referencesApplies dual-reference conditioningGenerates the face-swapped resultPreserves original scene compositionNo additional setup is required.What This Workflow PreservesOriginal backgroundOriginal compositionOriginal lightingOriginal perspectiveOriginal poseOriginal framingOriginal camera distanceOriginal scene layoutIdeal Use CasesUGC Creator ReplacementInfluencer Face SwapsCharacter ConsistencyMarketing CreativesAI AvatarsPersonal BrandingDataset GenerationBatch Face Swap PipelinesPerformanceTested on:NVIDIA T4RTX 3060RTX 4070RTX 4090VRAM requirements will vary depending on image resolution.Batch AutomationThis public release contains the single-image workflow.Internally, I use a custom Python automation system connected directly to the ComfyUI API that can process large batches automatically.Example setup:1 source face400+ target imagesAutomatic queue managementAutomatic retriesOrganized output foldersGPU-friendly processingDesigned for T4 cloud instancesThis automation system is not included in the public release.Commercial ApplicationsThis workflow can be used for:UGC generationMarketing campaignsCreator replacementCharacter consistency projectsAI influencer pipelinesContent automationNeed Batch Processing?I also develop custom automation systems built around ComfyUI.Examples:Face swap automationFolder-based processingComfyUI API integrationsLarge-scale image generation pipelinesDataset creation workflowsCustom workflow developmentIf you're interested in automation or custom solutions, feel free to send me a direct message on Civitai.CreditsBuilt and tested using FLUX.2 Klein 9B, latent reference conditioning, and a dedicated face swap LoRA for high-consistency identity transfer.If you find this workflow useful, consider leaving a rating and sharing your results.
FLUX.2-Klein-GGUF-Multi-LoRA-dAIver-v1.5 – Optimized Low-VRAM Multi-LoRA Workflow for Flux.2 Klein 9B (GGUF)
Optimized Low-VRAM Multi-LoRA Workflow for Flux.2 Klein 9B (GGUF)A refined and enhanced workflow by Experimental_dAIver and adapted for Flux.2 KleinThis workflow delivers the full power of Flux.2 Klein 9B in GGUF format, specially optimized for GPUs with less than 8 GB VRAM, like my RTX 4050 with only 6 GB. The integrated CacheDiT_Model_Optimizer and PatchSageAttentionKJ (both optional!) provide a noticeable speed boost with almost no quality loss. Full NdSuperLoraLoader support for easy multi-LoRA stacking with automatic trigger-word integration, the intuitive selectLatentSizePlus aspect-ratio and resolution selector, plus a complete SEEDVR2 Video Upscaler Subgraph for stunning high-resolution results complete this elegant and highly flexible setup.Version 1.5 is my initial public release and brings significant improvements in speed, usability, multi-LoRA handling and upscaling quality — while remaining extremely VRAM-efficient (tested on RTX 4050 with only 6 GB).Key features in v1.5:Full adaptation to Flux.2 Klein 9B with dedicated GGUF loader setup and the new Qwen3-8B GGUF text encoderNdSuperLoraLoader with full multi-LoRA support, auto-fetch trigger words and easy strength controlselectLatentSizePlus — intuitive aspect-ratio and resolution selector with beautiful presets (including golden-ratio-friendly options) plus easy orientation swapFull SEEDVR2 Video Upscaler Subgraph — powerful DiT-based high-end upscaler that delivers stunning 4K+ results with intelligent resolution handling, Lab color correction and temporal settings. Works exceptionally well on still images tooImproved workflow organization, expanded notes with recommended settings, and more robust saving options via SaveImageExtendedCacheDiT and SageAttention integration for optimized performance on the smaller, faster Flux.2 Klein modelRequired Custom NodesComfyUI-GGUF (UnetLoaderGGUF + CLIPLoaderGGUF)ComfyUI-CacheDiT (CacheDiT_Model_Optimizer)comfyui-kjnodes (PathchSageAttentionKJ and related nodes)ComfyUi-MzMaXaM (selectLatentSizePlus)save-image-extended-comfyui (SaveImageExtended)ComfyUI-SeedVR2_VideoUpscaler (full upscaler subgraph)Models & Downloads (exact paths)The following list explains the base models I am most frequently using with this workflow. The list as well explains where to put each file after you downloaded it.1. Main Model (Flux.2 Klein 9B GGUF):File: flux2Klein-9B-Q4_1.gguf (or any other GGUF quant you prefer)Download: Community GGUF quants for Flux.2 Klein are available on Hugging Face (search “flux2 klein gguf” or check repositories by city96 and similar quantizers)Target folder: ComfyUI/models/diffusion_models/ (or Flux/ subfolder)2. Text Encoder (CLIP)File: Qwen3-8B-Q4_K_M.ggufDownload: https://huggingface.co/mradermacher/Qwen3-8B-Q4_K_M-GGUF/resolve/main/Qwen3-8B-Q4_K_M.gguf (or equivalent Qwen3-8B GGUF quant)Target folder: ComfyUI/models/clip/ or ComfyUI/models/text_encoders/3. VAEFile: full_encoder_small_decoder.safetensors (specific VAE for Flux.2 Klein setups)Download: Usually provided together with Flux.2 Klein GGUF packages or available from the ComfyUI community / workflow author collectionTarget folder: ComfyUI/models/vae/4. Upscalers4× Upscaler: 4x-UltraSharp → https://civitai.com/models/116225/4x-ultrasharp1× Skin-Contrast Upscaler: 1xSkinContrast-High-SuperUltraCompact.pth → https://huggingface.co/notkenski/upscalers/blob/main/1xSkinContrast-High-SuperUltraCompact.pthTarget folder: ComfyUI/models/upscale_models/5. LoRAs (for NdSuperLoraLoader – full multi-LoRA support)Any Flux-compatible LoRA (.safetensors)Target folder: ComfyUI/models/loras/6. SEEDVR2 Models (for the high-end upscaler subgraph – optional but recommended)DiT Model: seedvr2_ema_3b-Q8_0.ggufVAE: ema_vae_fp16.safetensorsDownload from the official ComfyUI-SeedVR2_VideoUpscaler repository or Hugging Face and place in the folders required by the custom node.Recommended SettingsSampler: EulerScheduler: Simple or BetaSteps: 4CFG Scale: 1~1.5Shift: 4–7Resolution: freely selectable via the selectLatentSizePlus node (many golden-ratio-friendly presets included, swap orientation with one click)How to use the workflowLoad the JSON in ComfyUI.Select your desired aspect ratio and resolution in the selectLatentSizePlus node.Enter your prompt in the Positive Prompt node (the NdSuperLoraLoader automatically handles trigger words from your LoRAs).Load one or more Flux-compatible LoRAs into the NdSuperLoraLoader.Generate. The workflow produces a high-quality base image.(Optional) Run the SEEDVR2 upscaler subgraph for beautiful high-resolution results on images or video.Use SaveImageExtended for full filename, metadata and folder control.You can bypass the upscaler group completely for fast testing. The workflow also works with regular .safetensors Flux.2 Klein checkpoints if you replace the GGUF loaders with the standard nodes (higher VRAM usage).This is a clean, fast and powerful base specifically tuned for the new Flux.2 Klein 9B model — perfect for quick iterations, multi-LoRA compositions and high-quality final outputs even on modest hardware.Huge thanks to WikkedAI for the original foundation and to the creators of CacheDiT, SageAttention, NdSuperLoraLoader, selectLatentSizePlus and SeedVR2 for making Flux.2 Klein so accessible on low-VRAM GPUs.If you have questions or want to share your results — I’m happy to hear from you in the comments. Enjoy the workflow!
Flux2 Klein 9b/4b simple workflow (TXT2IMG, IMG2IMG, Edit, Inpaint, metadata export for Civitai)
A simple workflow for Klein family models in the first row was developed for 9b and 4b distilled versions, but it should work for base models with minimal changes (such as replacing ConditioningZeroOut with negative conditioning).v3.3 Latent This is an experimental version of the workflow. It’s not intended for general use (though it might still work as a regular one). The main goal here is to slow down degradation when editing images individually by one-by-one editioning. Typically, using VAE decode artifacts appears after 3–5 edits, which can be addressed by simply removing the latent encoding/decoding steps via two nodes: Latent Loader and Latent Save. This approach fully bypasses VAE entirely for preview purposes—only to use it during editing previews.I implemented a chain of K-samplers and used a specific schema; while there are no visible artifacts, over time the colors fade slightly, but this is still better than the standard Edit mode. If you try it out, please provide feedback about this workflow (whether positive or negative). I recognize that other methods exist for restoring images, but they’re much more complicated—so I’m exploring this approach first.At least for now, changing the size in Latent Edit mode isn’t available. It’s possible, but it comes with many limitations; I didn’t find the way I like. So you’ll have the same size as what was saved (.latent file).
My 4-In-1 Klein Workspace
Pheeby's 4-In-1 Flux Klein WorkspaceWhat this is and what it does:I've assembled my most often-used workflows into a single "workspace" file, because it saves me having to open separate tabs for different tasks and keep track of all of them, and eliminates weird bugs like LoRAs from different workflows merging and screwing up image generations and edits, or whatever other funky things can go wrong when you have multiple workflows open at the same time.The workspace includes sections for image generation, compositing, editing, and upscaling. The first section is used for creating images out of nothing but text. The next section is used for compositing these images into new images and editing them together into a seamless whole. The third section is used for various minor editing tasks like changing the style or lighting, color correction, etc, and finally, the last section is used for upscaling to achieve a finished result.It's been really convenient and useful for me to have all of these tools together in one place, so I'm sharing this on the off chance that someone else will find at least some part(s) of it useful as well, if only in constructing their own more ergonomically-designed-for-themselves workflow(s).There are some custom nodes required for this to be of any use to anyone, but there's nothing all that exotic about any of it, the nodes are actually useful and are not poo-poo nodes made of cockadoodle, and the Manager fixes you right up in no time, if you've ever used it you'll know this already, blah blah blah zzzzz.Nodes you'll need to use everything as-is:rgthree-comfy (https://github.com/rgthree/rgthree-comfy)Pixaroma (https://github.com/pixaroma/ComfyUI-Pixaroma)JPS Custom Nodes for ComfyUI (https://github.com/JPS-GER/ComfyUI_JPS-Nodes)ComfyUI-ReferenceChain (https://github.com/remingtonspaz/ComfyUI-ReferenceChain)ComfyUI-GGUF (https://github.com/city96/ComfyUI-GGUF) <-- if using GGUF modelsNaturally, you will want to change some things around to make it all work for you, especially if you're not using GGUF models, but all of this setup works fine with regular models and with the 4B variant of Klein, you'll just have to swap out a couple of nodes. Refer to the README file if/when you get mysterious errors for no apparent reason. As always, YMMV, good luck, and have fun~!PS: I do very graciously accept tips, but I don't like annoying people by asking them for anything, because it just feels weird. Just know that if people tip me, I will have the resources to try my hand at training. I don't trust my 8GB VRAM to allow me to finish training a LoRA at home before I die of old age. XD
Flux Klein 9B 3 Image Edit
A simple 3-image Klein 9B edit workflow for ComfyUI. This workflow is built for combining multiple image references into one edited output using a prompt-driven FLUX/Klein setup. Drop in three source images, describe how you want them combined, and let the workflow handle the reference conditioning, edit pass, detail refinement, upscaling, and final sharpening.The setup uses reference latents from all three input images, a two-stage sampling flow, Detail Daemon refinement, Remacri upscaling, resize to 1920px longest side, and CAS sharpening for a cleaner final image.Good for quick composites, character/object placement, car edits, scene edits, and testing multi-image reference control without building a huge complicated workflow from scratch.This is a basic V1 workflow, so treat it as a starting point and tweak CFG, denoise, steps, sampler, and prompt strength depending on your images. This workflow uses res2 sampling.
FLUX 2 KLEIN 9B — Smart Character Replacement Workflow
This workflow focuses on precise character replacement inside crowded or complex images using ComfyUI native nodes only.The main goal was simple:replace one specific person from a group using a reference image — without destroying the rest of the image quality.
Flux 2 Klein Basic with Reference Image
This is a very simple ComfyUI workflow for creating Flux 2 Klein images using a reference image, for those who are newer to ComfyUI or Klein. The only custom node included is the Clip model load from https://github.com/city96/ComfyUI-GGUF so that you can find it via missing nodes in the manager if you want to use GGUF files. Then you can swap out the model loader and clip loader with GGUF / non-GGUF versions as you please.
Goon's - F2K I2I Edit + SeedVR2 Upscaling
- Goons F2K I2I + SeedVR2 -High quality Flux2 Klein 9B image-to-image editing workflow focused on simplicity and being slightly more beginner friendly, cleaner generations, and improved quality through up-scaling while also ensuring detail preservation and remaining lightweight enough for lower-end hardware setups.Now updated with SeedVR2 upscaling for significantly improved final output quality without the usual over-sharpening, artifacting, or unwanted / un-needed hallucinated details common with many upscale workflows.✨ New in v1.2Added SeedVR2 upscaling to even further increase output quality without any un-wanted detail loss/adding.change from default model to "Moody Desire Mix - V2.0 EDIT + HEADSWAP" from @catlover1937 for MUCH better nsfw support/quality.seperated image loading nodes that were previously packed into subgraphs with image loaders to ensure using maskeditor on the load image node is still possible.Runs much easier on lower hardware (12gb+ vram reccomended)🔥 FeaturesFlux2 Klein 9B powered image-to-image editingStrong prompt adherenceHigh detail preservationSeedVR2 integrated upscalingBetter NSFW supportMask Editor compatibilityOptimized for quality-to-performance balance🖥️ Hardware Requirements12GB+ VRAM recommendedLower VRAM setups may still work depending on resolution and settings.Designed for users who want high-end edit quality without requiring extreme hardware.Please feel free to dm me with any questions, requests, suggestions, issues, or pretty much anything and ill repsond as fast as i can :)
Flux 2 Klein 9B destiled Easy to use, Lora Manager + Triggers (By: Rafaelldestilo)
I'm sharing my workflow that I use for basically any task.It features easy image aspect activation; just select the one you want.Sage Attention is activated for quick generation; if you don't have it, just deactivate it.Lora Manager - where you can store all your Loras; hovering the cursor over them shows a cover image from the store, greatly helping with style identification. When activated, it pulls all activation keys for easy use, eliminating the need to search for activation keys, as it's directly synchronized by Civitate.It's a straightforward, easy, and simple workflow with high-resolution image generation and very fast speed.Lora manager downloadhttps://github.com/willmiao/ComfyUI-Lora-Manager
Klein_9b Pro Grade Face/ Clothes/ Item Swap
I got tired of Swaps or inpainting that always looked like cardboard cutouts, or the swapped out area color was horrible.This is for Klein_9b only. This does not work with other Klein modelsThis is a combination of a group of nodes that accomplish several things that a other workflows I have seen here have not been able to do:Better reference latent balancingBetter conditioning text reference balancingReference latent controllers that allow you to balance insted of "hoping for the best"COLOR correction across both images for insane consistency instead of smashing one with the other.Upscaler for higher quality images🎨 Post production Suite for final adjustmentsI made it modular so you an strip out what you need and use it in your own workflows.Note: If you have any issues with this workflow or questions, please feel free to message me or drop a note in the comments.Instagram: https://www.instagram.com/synth.studio.models/Buy me a☕ https://ko-fi.com/lonecatoneThis represents many of hours of work. If you enjoy it, please 👍like, 💬 comment , and feel free to ⚡tip 😉
RealityFlow
this is the place where you can find my experiments with ultimate photorealism ...
Scene character replacement（flux2 klein）
More workflow parameter settings, please refer to my settings on rh:https://www.runninghub.ai/post/2053851228840439809/?inviteCode=rh-v1315Bind my invitation code: rh-v1315 to receive an extra 1000 RH coins. You can also earn 100 points by logging in daily, which can be used for generating images or videos.My AI models/apps are on RH now! Try them out & support my work — every click counts. Thank you!Thank you all for your support. Your support is the motivation for me to update more interesting and useful workflows. Thank you so much!!!Models used in the workflow:diffusion models:https://huggingface.co/wikeeyang/Flux2-Klein-9B-True-V1/tree/mainhttps://huggingface.co/wikeeyang/Flux2-Klein-9B-True-V2/tree/maintext encoders:https://huggingface.co/Comfy-Org/vae-text-encorder-for-flux-klein-9b/tree/main/split_filesloras:https://civitai.red/models/2191265/mannequins-pose-references-qwen-image-edit-2509-flux-2-klein-9b-base?modelVersionId=2689061https://civitai.red/models/2027766/bfs-best-face-swap?modelVersionId=2610018https://civitai.red/models/1939453/consistence-edit-lora?modelVersionId=2863285https://civitai.red/models/2436859/high-resolutionhttps://civitai.red/models/2605373/kleinanime-hd-upscaling?modelVersionId=2925488checkpoints:https://civitai.red/models/827184/wai-illustrious-sdxlLLM:https://huggingface.co/HauhauCS/Qwen3.5-9B-Uncensored-HauhauCS-Aggressive/tree/mainCustom nodes used in the workflow：https://github.com/Kosinkadink/ComfyUI-Advanced-ControlNethttps://github.com/yolain/ComfyUI-Easy-Usehttps://github.com/rookiepsi/comfyui-extendedhttps://github.com/iacoposk8/ComfyUI-Fooocus-Inpaint-Wrapperhttps://github.com/ltdrdata/ComfyUI-Impact-Packhttps://github.com/ltdrdata/ComfyUI-Impact-Subpackhttps://github.com/kijai/ComfyUI-KJNodeshttps://github.com/lihaoyun6/ComfyUI-llama-cpp_vllmhttps://github.com/NewLouwa/ComfyUI-Model_preset_Pilothttps://github.com/1038lab/ComfyUI-RMBGhttps://github.com/Suzie1/ComfyUI_Comfyroll_CustomNodeshttps://github.com/audioscavenger/ComfyUI-Thumbnailshttps://github.com/chflame163/ComfyUI_LayerStylehttps://github.com/brucew4yn3rp/ComfyUI_SelectiveMetadatahttps://github.com/aining2022/ComfyUI_Swwanhttps://github.com/Pirog17000/Pirogs-Nodeshttps://github.com/rgthree/rgthree-comfyhttps://github.com/JamePeng/llama-cpp-python/releases
Face Consistency Workflow ( flux Klein Enhancer )
I have made the video on this take a look if you want to know more
Flux Klein 9b Scalable Multi-run w/ T2I and Inpainting Toggles
OverviewA workflow for Flux Klein 9 designed for easy scaling of reference images and runs. It is multi-run so your resulting image will be pushed into the next run allowing you to iterate on it. It also includes Text-to-Image and Inpainting Options. You will see the options for each included on the "Run" subgraphs. Ok, I am stretching the word "scalable", but you can add in how ever many additional runs as you want. You can also go into the subgraphs and add more reference images if you desire. YOU MUST HAVE AN IMAGE LOADED IN EACH LOAD IMAGE SLOT. The work flow takes advantages of switches instead of bypassing subgraphs. I think it makes it easier, but the side effect is if you don't have images loaded for each reference (even if it's ignoring them) the run will fail. I'm not going to go over the directions as I think it's pretty straight forward if you've used FK9 before. If you want me to go into detail, though, or update the workflow to include notes let me know.
Create Faces From 3-Images (Unique Face Creator)
About Sam3.1 I Upgrade wf to native sam nodes. Download and put it to model/checkpoint https://huggingface.co/Comfy-Org/sam3.1/tree/main/checkpointsAbout V2; Only python 3.12. If anyone runs this node on Python 3.13, can they share it with me?I gave up on the problematic original Sam3 node and switched to the "Easy Sam3" node***Please upgrade your photobuf and mediapipe for "personmask" ( python_embeded\python.exe -m pip install --upgrade mediapipe protobuf google-cloud-videointelligence )***Create Your Influencer Face From 3-Images.Yes, this time it worked, and I wanted to dedicate a separate section to it. In my previous workflow, I created a face from two images but couldn't quite achieve what I wanted. This time, I was truly successful in creating a unique face.This is a simple workflow for generating new, unique faces. It uses SAM3 to mask specific features (like the nose or mouth) from a source image and then builds a completely different face around those parts. Good for creating character variations or unique identities. You can freely decide which parts of the face to mask. Simply write it in the prompt of the Sam3 model. Keep in mind that some faces are quite dominant, and the model may draw too much inspiration from them. You may need to apply more masking (restrictions).Why did I need it? Because Pulid and Reactor weren't satisfying me in this regard. Face ID was creating unrealistic faces. Is this workflow more successful than those? You decide. Sometimes AI when trying to fit the mouth and nose into masked areas, a situation different from the reference ones may arise. IDK. Maybe someone will be inspired and develop it further in the future.How it works: SAM3 Masking: Automatically selects facial features to keep as a reference.Generation: it tries to create a new person using the unmasked parts.Quick Guide:Upload your source image.Use the SAM3 prompts to choose which parts to keep.If you like my work, you can buy me a coffee!☕ https://buymeacoffee.com/ugurdoyduk
Remove Object From Image Workflow Bbox Flux Klein-9B GGUF
🏃AboutA simple workflow for object removal tasks. Uses red bounding boxes (bbox for short). Runs on the Flux 2 Klein 9B Q4_K_S GGUF model and object removal lora.💻My specs🔥GPU - RTX 5070 12GB📦RAM - 48 GB RAM DDR5 6000🧠CPU - Ryzen 7500F🌐Model linksMain model FLUX.2-klein-9B-GGUF: https://huggingface.co/unsloth/FLUX.2-klein-9B-GGUFObject removal lora: https://huggingface.co/prithivMLmods/FLUX.2-Klein-Object-Remover-Bbox📖How to use1) Make red rectangles on your image 2) Load clip, vae, gguf model, lora3) Run
f2k9b One-Click Workflow to Real Person Conversion (No LoRa Required)
f2k9b One-Click Workflow to Real Person Conversion (No LoRa Required)I've seen many workflows for converting to real person conversion are very complex, some even requiring LoRa. After testing, LoRa doesn't seem to work much, or rather, the only thing that really matters is the "Reference Latent" node. Some workflows don't control the ratio, making it impossible to adjust the redraw magnitude.I prefer a minimalist approach, using almost all the nodes that come with ComfyUI. You can replace gguf with safetensors, and replace the image saving node with the built-in one; everything else remains unchanged.I don't know how to upload the workflow, so I've included images at the end.f2k9b 一键转真人工作流（无需lora）因看到很多转真人的工作流很复杂，有的还要lora。经测试，lora没啥效果，或者说真正起作用的是“参考Latent”这个节点。有的工作流没有控制比值，无法调节重绘幅度。本人喜欢极简，几乎就是comfyui自带的节点。gguf可以替换成safetensors，保存图像节点换成自带的，其他的不用变。
ControlNet depth or openpose flux2klein
Version v1.2:Latend Select (optional) PulIDFlux2 Custom SamplerVersion v1.1:You can now choose between four options via Option Select.Option 0 - Aio Aux (ComfyUI Controllnet Aux)Option 1 - Manually Openpose (comfyui-openpose-editor-konva)Option 2 - Automatic Pose (DWpose)Option 3 - Advanced Automatic Pose 2 (westNeighbor / ComfyUI-ultimate-openpose-estimator)So far, everything is relatively self-explanatory.Version v1.0This workflow gives you complete control over your character's pose. Tested with Flux2klein.Using the CR Image Input Switch, you can choose between depth or open pose.Alternatively, instead of depth or open pose, you can connect a Load Image to the CR Image Input Switch and then load an image with a drawn pose.
All-in-One. Z-Image, Ernie, Klein, t2i, i2i, controlNet, inpaint, outpaint, LLM, WAN, Qwen image, SDXL, Chroma
!!! Before installing custom nodes from manager you need to install this one from Git urlhttps://github.com/BigStationW/ComfyUI-ppm - negative prompts for Flux1, ZIT and AnimaFor example if you don't want to see glasses and tie on Anima, Zit and Flux1 generations add to fixed prompt window string (glasses, tie:-1)If manager fails to find LLM and lora nodes, install them from git url - https://github.com/willmiao/ComfyUI-Lora-Managerhttps://github.com/KingManiya/ComfyUI-LLM-text-processor (Automatic setup currently supports Windows x64 CUDA 13 only. Other platforms require manual setup of llama.cpp release binaries.)Since v.8.3 there is a separate version of workflow that can be used with Sage Attention enabled in Comfyui launch arguments (it can be easily installed from https://github.com/wildminder/AI-windows-whl?tab=readme-ov-file with no need to install compiler)This workflow supportZ-Image, Ernie image, MS Lens, Wan, Flux, Flux 2, Flux 2 Klein, Anima, Qwen Image, Qwen Image Edit, SDXL/Pony, Chroma and Lumina-Image 2.0txt2img, img2img, Inpaint functionality.Face swap (not that perfect atm)SeedVR2 upscalerLLM to describe images and enhance promptsControlNet only for Edit models in img2img group (just describe your image in prompt and add some instruction like "use input image as normal map reference")Outpainting in Edit/Kontext modesafetensor, gguf and svdq (Nunchaku) checkpointsPowerful multi Lora loader (Lora manager)Face After DetailerMulti or single images load for Edit/KontextLLM Models:VRAM 8 Gb:Qwen3.5-4B-Uncensored-HauhauCS-Aggressive-Q8_0.ggufmmproj-Qwen3.5-4B-Uncensored-HauhauCS-Aggressive-BF16.ggufVRAM 12+ Gb:Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q4_K_M.ggufmmproj-Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-f16.gguforGemma-4-E4B-Uncensored-HauhauCS-Aggressive-Q6_K_P.ggufmmproj-Gemma-4-E4B-Uncensored-HauhauCS-Aggressive-f16.ggufPut models and folder with system prompt here:📂 ComfyUI/ ├── 📂 models/ │ └── 📂 LLM/ │ ├── Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-Q4_K_M.gguf │ ├── mmproj-Qwen3.6-35B-A3B-Uncensored-HauhauCS-Aggressive-f16.gguf │ └── 📂 prompts/ │ ├── LLM-system.txt │ ├── LLM-system-SDXL.txtStarting from v8.0, it is possible to enhance T2I prompts on the fly using switches in the T2I group. You still need to manually copy and paste prompts from the LLM group to the I2I group, but you can directly describe Image 1 loaded in the I2I block.Follow the short block notes for more information.User prompts in both groups are now divided into two parts. The first part (Fixed) is never passed to the LLM block for processing and will always remain unchanged at the beginning of the prompt. The second part (User Prompt) can be passed to the LLM (if enabled) and improved.
Flux.2 Klein Mix Workflow. Consistency, Enhanced, Image Size control, I2I, T2I, In & outpaint, ControlNets.
Flux.2 Klein Mix Workflow Workflow with main focus on Image Editing, with control over Image Sizes. Trying to only use the more common Custom Nodes or ones that do not have requirements for a easy and safe installation. Aiming for consistency and support edits for up to 4 images and Text to Image, 3 different Controlnets, different Inpainting methods and Outpainting. Nodes for selecting image sizes and Aspect Ratio. Uses Flux2 Klein Enhancer nodes, for consistency and prompt impact. Color matching and different sampler with ETA setting.The workflow has many notes that explain things inside, which you can read or delete.There is a second Version inside for people that use Lora Manager so you can pick what you want.The workflow only support a Positive Prompt except for the 2-pass mode in V5 for T2I that uses both the Base and the Normal model and a negative prompt.All custom nodes used:EssentialsRgthreeEasy-UseImpact PackKjNodesGGUFControlnet_auxLayerStyleArtVentureBelow Does not have any requirements making for an easier and safer install:Inpaint-CropAndStitchModdedNodesComfyrollControlAltAIFlux2Klein-EnhancerCRZnodesmxToolkitZImageTurbo-FlowSamplerNote that V2 and V1 do not use the same custom nodes, and are harder to use because of that, and I do not recommend using them!Feel free to criticize, comment or ask questions.
🎬 The Great Escape Flux (Beta)
🎬 The Great Escape Flux (Beta)Ever wanted to drop your friends, celebrities, or a random selfie into a high-stakes action movie where they’re desperately sprinting from T-Rexes, colossal spiders, angry geese, rogue AI robots, or a tidal wave of cute kittens?This node is your personal meme director. It generates dynamic, cinematic escape scenes using Flux, giving you full control over the threat, emotion intensity, camera framing, and cinematic style—while doing its absolute best to keep the original face recognizable.⚠️ Beta / Experimental ToolGenerative AI is beautifully unpredictable. Please treat this as a creative playground, not a photocopier. Results will vary depending on:🔢 Seed & Flux randomness – AI rolls the dice every time.📸 Source image quality – Clear faces, good lighting, and frontal angles work best. Extreme poses or heavy shadows may cause drift.🌪️ Scene complexity – Some wild combinations (e.g., Close-Up + Extreme Emotion + Kaiju Attack) may produce artifacts, extra limbs, or surreal physics. That’s part of the charm.🎭 Important Usage Notes👤 Identity Preservation: Works strongest on Close-Up and Medium Close-Up shots. The wider the frame, the more Flux may "reinterpret" facial features. Don't be afraid to experiment with preserve_identity and preserve_clothing toggles.🎚️ Emotion Strength: Use the strength_level dropdown to dial intensity from "Minimal (trace of)" to "Extreme (overwhelming)". Higher strength + wider shots = higher chance of identity drift.💡 Pro Tip: Generate 3–5 variants and pick the best one. AI is a collaborator, not a factory. Slight seed changes often turn a "meh" into a meme legend.📦 InstallationDownload the release archive (contains the ComfyUI-Flux Meme Director folder + a simple example workflow).Move the ComfyUI-Flux Meme Director folder into your ComfyUI/custom_nodes/ directory.Restart ComfyUI (or click Refresh in the manager).Right-click → Add Node → Search for 🎬 The Great Escape Flux (Beta) or load the included workflow to get started instantly.🤝 Community & FeedbackThis is a community-driven beta. I’d love to see what you create!🖼️ Share your funniest, most epic, or most cursed generations.🐛 Report bugs, weird artifacts, or broken combinations.💡 Suggest new chase scenarios, emotions, or cinematic styles.Your feedback directly shapes the next update. Let’s make escaping from digital threats an art form.🏃♂️💨 Run. Generate. Survive.
Moody Simple Flux 2 Klein 9B Edit Workflow
I just made a new Telegram Group and Discord group drop in for chats, questions, feedbacks or just to share your work. 欢迎加入纸飞机和Discord群组。喜欢我的作品吗？🎨 我会持续免费分享更多优质内容✨，欢迎支持一下，请我喝杯咖啡吧！💖 Like my work? 🎨 I share all my work for free ✨, so support me by buying me a coffee 💖!Ko-fi: ko-fi.com/catlover1937Fanvue: https://www.fanvue.com/catlover1937Please do not redistribute without consent. 如非作者同意禁止搬运。2026-05-03V2.0例图有工作流例子。 Examples in demo photos.QoL updates and better fit for Moody Desire Mix V2.0.体验感更新 + 更好兼容 Moody Desire Mix V2.0.2026-04-06Simple workflow for image editing! ✨The demo images were generated using Moody Desire Mix.Please note that the workflows shown in the images may not be the final version, so be sure to download the latest workflow from the model page. 📥简单易用的图像编辑工作流！✨示例图片由 Moody Desire Mix 生成。请注意，图片中展示的工作流可能不是最终版本，请务必前往模型页面下载最新版本。📥
Moody Simple Flux 2 Klein 9B Workflow
I just made a new Telegram Group and Discord group drop in for chats, questions, feedbacks or just to share your work. 欢迎加入纸飞机和Discord群组。喜欢我的作品吗？🎨 我会持续免费分享更多优质内容✨，欢迎支持一下，请我喝杯咖啡吧！💖 Like my work? 🎨 I share all my work for free ✨, so support me by buying me a coffee 💖!Ko-fi: ko-fi.com/catlover1937Fanvue: https://www.fanvue.com/catlover1937Please do not redistribute without consent. 如非作者同意禁止搬运。2026-05-01🚀 V2 Release NotesHi everyone 👋I’ve updated this workflow to V2.✨ What’s New🔧 Process improvementsRefined the workflow based on a better understanding of F2K generation behavior.🧠 Improved 2nd pass logicThe second pass now only enhances the first image. Since F2K is an edit model, running two full passes (as in V1) was not ideal — this has now been corrected.🎯 Better support for Moody Desire V2 (LITE + PRO)Adjusted for improved compatibility and performance with both versions.⚡ Quality-of-life improvementsGeneral usability improvements for a smoother workflow experience.V2 更新说明大家好 👋这个工作流已经更新到 V2。✨ 更新内容🔧 流程优化基于对 F2K 生成机制更深入的理解，对流程进行了改进。🧠 优化二次处理逻辑第二次处理现在只用于增强第一次生成的结果。由于 F2K 是编辑模型，像 V1 那样进行两次完整生成并不合理，这一点已经修正。🎯 更好支持 Moody Desire V2（LITE + PRO）提升了对两个版本的兼容性和表现。⚡ 体验优化（QoL）提升整体使用体验，让流程更加顺畅。2026-04-06Simple workflow for image generation! ✨The demo images were generated using Moody Desire Mix.Please note that the workflows shown in the images may not be the final version, so be sure to download the latest workflow from the model page. 📥简单易用的图像生成工作流！✨示例图片由 Moody Desire Mix 生成。请注意，图片中展示的工作流可能不是最终版本，请务必前往模型页面下载最新版本。📥
Klein 9B img2img (denoise) with Loras
Klein 9B img2img (denoise) with Loras
Klein 9B Edit with Loras
Klein 9B image edit workflow for Comfyui with Loras support. CTRL + B to activate Lora Node.
F2K9BCompact
New to comfy, sharing my workflow for criticism and editing.
A Advanced Workflows for Capitan01R and all users - Enhancer - Prompts gestion
Hello and thank you to Capitan01R for their Nodes! I'm presenting a workflow created using some of their Nodes.V3b: Hello! I’ve made a few improvements to the model management system. Unfortunately, I had to remove three workflows and replace them with blank placeholders — which you’re welcome to fill with your own work or any workflow that serves you well.The reason:First, I didn’t receive responses from users confirming whether I could safely integrate their workflows into this publicly shared template (with proper attribution).Second, I wasn’t able to locate the original creators of severals workflows.As a result, I made the difficult decision to publish only the version that best preserves neutrality and fairness — while still leaving space for your contributions.That said, I sincerely thank all the creators whose work may have been included anonymously—I truly appreciate your creativity and generosity, even if I couldn’t credit you this time aroundI hope it will be useful to you. I've also included a link to their account where you can find more information about their work. https://civitai.red/models/2492746?modelVersionId=2849381The main focus of this workflow is to enhance image quality, and I've added a system for evolving prompts that can serve as a foundation for beginners.I’m not very strong in English, so there may be times when some prompts don’t do exactly what I intend.Please feel free to correct and refine them — and thank you for your understanding!Best regardsA note: remember to specify the save folder for your prompt files.edit : oupss i forget to write : Thanks to all creators and have all funrious ;-)
Random Prompts Workflow
Create a random prompt based on the options provided.Place the options between {} separated by |.example {gray|back|blue|white}It works with any model {SD|Flux|Wan}...etc.
Fix that Image! (SFW/NSFW) Klein_9b version
Repaired after updateNote: If you have ANY issues with nodes not downloading, read the notes or reach out. There's nothing that special about any of them that aren't core modules.⛔⚠️🛑✋ Read the notes completly before using. Most common install and node problems are listed in the directionsA troubleshooting guide including a list of common error messages can be found here https://civitai.com/articles/24678/how-to-use-my-workflows-a-comprehensive-breakdown"Man, if only....."Need better resolution? Got it.Want to change her clothes? Sure.What her naked? No problem.want her dressed? Umm...okay.Workflow features:InpaintingHi rez fixSeed VR2 UpscalerMy exclusive Post Processing SuiteInstagram: https://www.instagram.com/synth.studio.models/Buy me a☕ https://ko-fi.com/lonecatoneThis represents many of hours of work. If you enjoy it, please 👍like, 💬 comment , and feel free to ⚡tip 😉
Moody Anime to Real Workflow
I just made a new Telegram Group and Discord group drop in for chats, questions, feedbacks or just to share your work. 欢迎加入纸飞机和Discord群组。喜欢我的作品吗？🎨 我会持续免费分享更多优质内容✨，欢迎支持一下，请我喝杯咖啡吧！💖 Like my work? 🎨 I share all my work for free ✨, so support me by buying me a coffee 💖!Ko-fi: ko-fi.com/catlover1937Fanvue: https://www.fanvue.com/catlover1937Please do not redistribute without consent. 如非作者同意禁止搬运。2026-04-28✨ Release Notes V3 — Stable Version🇺🇸 EnglishHi everyone! 👋As promised, I’m finally bringing you a fully functional Anime → Real workflow — this is no longer an experimental setup, but something you can actually rely on for consistent results. 🚀🔥 What’s New✅ Stable & reproducible pipeline — tested and confirmed to work consistently🎯 Designed specifically for anime → realistic character transformation⚡ Improved realism, skin texture, lighting, and facial structure fidelity🧠 How It WorksThis workflow is built around my F2K model — Moody Desire Mix 💎It acts as the core transformation model, converting anime features into realistic human structureHandles base identity, proportions, and overall style translationHowever, to push realism even further… 👇✨ Enhancement Stack (Optional but Recommended)Paired with my ZIT models to:Enhance photorealismImprove detail richnessRefine lighting & texture quality💡 You can skip the ZIT stage if you prefer a lighter or faster workflow, but for best results, I highly recommend keeping it enabled.🎭 Character RealismIncluded in the ZIP: Anime → Real consistency LoRAHelps maintain realistic character across generationsHuge thanks to @AIGC_Singularity for his lora contribution 🙏📦 Package IncludesFull workflow (ready to import)Required LoRA (pre-configured)Example settings for quick start⚠️ NotesWorks best with high-quality anime inputsResults may vary depending on pose, lighting, and prompt qualityGPU memory requirements can vary depending on upscale settings💬 Final WordsYour support really means a lot ❤️I’ve been getting pretty busy with life lately, but I’ll continue doing my best to improve and update my models and workflows whenever possible.Hope you enjoy this one — and as always, feel free to share your results! 🔥Cheers 🍻🇨🇳 中文大家好！👋正如之前承诺的，这次带来的是一个真正稳定可用的 Anime ➜ 真人工作流，不再是实验性质，这次是可以放心用来出图的版本了 🚀🔥 更新内容✅ 稳定可复现流程 — 已测试，基本可以保证成功运行🎯 专门针对 二次元转写实人物 优化⚡ 大幅提升真实感（皮肤、光影、五官结构）🧠 工作原理本工作流核心使用我的模型：F2K — Moody Desire Mix 💎作为主转换模型，负责从动漫风格转为真人结构控制整体人物比例、面部结构和风格过渡但仅靠它还不够极致，所以加入了👇✨ 增强模块（可选）搭配我的 ZIT 系列模型：提升照片级真实感增强细节表现优化光影与纹理💡 如果你想更快或更轻量，也可以跳过这一部分，但建议开启以获得最佳效果。🎭 角色真实性压缩包中已包含：Anime ➜ Real 一致性 LoRA用于保持角色生成中的稳定性与真实性特别感谢 @AIGC_Singularity 提供的Lora支持🙏📦 包含内容完整工作流（可直接导入）所需 LoRA（已配置在ZIP文档内）示例参数，方便快速上手⚠️ 注意事项建议使用高质量动漫输入图效果会受到姿势、光照、提示词影响显存占用取决于放大与增强设置💬 最后想说真的很感谢大家的支持 ❤️最近现实生活比较忙，但我还是会尽量持续更新模型和工作流！希望你们会喜欢这次的内容，也欢迎分享你们的成果 🔥干杯 🍻V3展示图：2026-04-21V2.0 Release！EN 🇺🇸⚠️ DISCLAIMER:This is NOT the “perfect” anime ➜ real workflow — it’s simply me pushing the limits of what Zimage ControlNet can do right now. Without a proper Zimage Edit model release, this will most likely be the final update for this workflow. I am more like testing the upper limit of Zimage without using any Edit models.That said… I’ll be working on a true anime ➜ real pipeline using my Flux.2 Klein model 👀so stay tuned.✨ What’s New:🔧 Adjusted ControlNet settings to reduce defects📏 Dynamic resizing based on long edge🧠 Dual ControlNet setup for better structure control✍️ Stronger Florence 2 prompt generation with more aggressive filtering🗂️ Cleaner, reorganized workflow + optimized settings (aligned with my newer workflows)🆚 Added side-by-side comparison preview + auto save💡 Added built-in tips for generation & troubleshooting🛟 QoL update: new “save me daddy” switch⚠️ What’s Bad:🎭 Results may look less similar to the original image⏳ Slightly longer generation time❤️ Final Words:Thanks for all the support along the way — really appreciate it 🙏Have fun and enjoy creating! 🎉CN 🇨🇳⚠️ 免责声明：这并不是“完美”的动漫 ➜ 写实工作流 —— 只是我在测试 Zimage ControlNet 的极限能力。在没有 Zimage Edit 模型发布的情况下，这大概率会是这个工作流的最后一次更新。我主要是想测试不用编辑模型的情况下，Zimage + Controlnet的极限。不过我最近将会基于我的 Flux.2 Klein 模型 做一个真正成熟的动漫 ➜ 写实方案 👀 敬请期待。✨ 更新内容：🔧 调整 ControlNet 参数，减少瑕疵📏 基于长边的动态分辨率调整🧠 双 ControlNet 模型，提高结构控制能力✍️ 更强的 Florence 2 提示词生成（更严格过滤）🗂️ 重构并优化工作流结构（对齐我最新的工作流风格）🆚 新增对比预览 + 自动保存💡 内置生成与排错提示🛟 体验优化：新增 “save me daddy” 开关⚠️ 不足之处：🎭 成图与原图相似度会降低⏳ 生成时间略有增加❤️ 最后：感谢大家一直以来的支持 🙏祝大家玩得开心，玩出好图！🎉Some examples，转换展示:2026-03-23This workflow is designed to transform anime-style images into high-quality, photorealistic outputs while preserving the original composition, character identity, and mood.Unlike simple style transfer, this pipeline focuses on reconstructing realistic facial structure, skin texture, lighting, and material details from stylized inputs. The goal is not just realism—but believable photography.🔹 Key FeaturesCharacter PreservationMaintains pose, framing, and key identity traits from the original anime image.Photorealistic Skin & LightingRebuilds natural skin tones, micro-textures, and physically plausible lighting.Controlled Stylization BalanceReduces anime artifacts (flat shading, linework) while avoiding overcorrection or distortion.High-Resolution FriendlyWorks well with upscaling pipelines for clean, detailed final outputs.This workflow also integrates Florence 2 for automatic prompt detection, along with Zimage Turbo ControlNet to preserve pose and composition. The settings are carefully tuned—please try to keep them unchanged, as they’ve been optimized over 3 sleepless days and nights to achieve a balanced result.The workflow is compatible with all Moody Zimage Turbo and distilled models. All demo images were generated using Moody Porn V10 DPO.该工作流旨在将动漫风格图像转换为高质量、照片级写实的输出，同时保留原始构图、角色特征以及整体氛围。不同于简单的风格迁移，这一流程专注于从风格化输入中重建真实的人脸结构、皮肤质感、光照效果以及材质细节。目标不仅是“看起来真实”，而是达到“可信的摄影效果”。🔹 核心特点角色保留保持原始动漫图像中的姿势、构图以及关键角色特征。照片级皮肤与光照重建自然肤色、微观纹理以及符合物理逻辑的光照效果。风格平衡控制减少动漫特有的痕迹（如平涂、线稿），同时避免过度修正或形变。高分辨率友好兼容放大流程，能够输出干净且细节丰富的最终图像。该工作流还集成了 Florence 2 用于自动提示词识别，并结合 Zimage Turbo ControlNet 来保持姿势与构图一致。所有参数均经过多轮精细调校（我测试+微调了三天三夜了，您就别折腾了，坏了又来跟我抱怨），请尽量保持默认设置不变，以获得最佳平衡效果。该工作流兼容所有 Moody Zimage Turbo 及蒸馏模型。所有演示图片均使用 Moody Porn V10 DPO 生成。FAQ:Still look kinda anime?Paste the output back to input and run again!结果不尽人意，还是偏动漫吗？把出的图复制粘贴回去，回炉重练一次！
Edit workflow 2 in one - Flux Klein 9 + Qwen edit 2511 + SAM3
Workflow OverviewThis is a dual editing workflow for Qwen Edit 2511 and Flux Klein 9B; however, both model groups are self-contained. You can safely delete the entire group of whichever model you don't use without breaking the workflow. The remaining model group will work independently without any additional changes needed.Additionally, the workflow can function as a standard text-to-image workflow.Why a dual workflow?Flux Klein is my preferred editing model and I use it all the time; however, I like to have the Qwen model as a backup when Klein struggles with certain edits, and vice versa.Some personal notes:Qwen Edit has better character consistency, but it loses skin details; Flux Klein can sometimes affect character consistency, but it preserves details better.Two variants are provided in the zip file:A standard edit workflow (EditStandard.json)The same workflow as the first one, but with the extra feature of saving images with metadata compatible with CivitAI (EditCivitAI.json)Main featuresStandard text-to-image generationEdit images guided by a text prompt, using up to three reference source imagesIsolate and mask specific objects or subjects via SAM3 segmentation or hand-drawn masks (or both) to avoid pixel driftingCamera angle control (especially useful with Qwen + multi-angle LoRA)Basic pose transferExample use casesVirtual try-on / outfit transferObject / subject removalBackground / scene changeCharacter swap / face swapInpaintingMulti-angle viewPose changeCharacter sheetScene compositionImage fixingImage style changeDetailed instructions are contained in the workflow itself:Red nodes are instructions and useful notes.Yellow nodes are configurable elements you can adjust to your needs.
Object Swapping flux-2-klein-9b
swap objects using a reference image. It’s pretty smooth - it uses SAM2 for the segmentation and SEEDVR to push the final result to 4K.
Flux Klein & Ollama | Ultimate AI Influencer Lookbook Creator (True Consistency)
About V2; Now this is much more faster. I added Ollama Llm instead of Qwenvl. Much faster. I changed and edited the Ollama-prompts. Download Ollama from Download Ollama on Windows. install to your pc. Then install missing nodes. Thats it."The only workflow you need for consistent AI Influencer content."Are you struggling to keep your AI Influencer’s outfit consistent across different poses and backgrounds? This ComfyUI workflow solves the most common "virtual try-on" issues by combining the anatomical power of Flux.1 with the visual intelligence of Qwen-VL.☕ Buy Me a Coffee:🔗 https://buymeacoffee.com/ugurdoyduk🚀 Key Features:High-Fidelity Extraction: Uses the Flux2 Klein method to isolate clothing from any reference image onto a clean white background, preserving every texture and detail.Smart Auto-Prompting (Qwen-VL): Forget about manual prompting! The integrated Qwen-VL LLM analyzes the scene and generates natural, aesthetic, and trendy poses for your influencer automatically.Face & Character Protection: Integrated Yolo Cropper ensuring your influencer's face remains 100% consistent and sharp, even during complex outfit swaps.Pro-Photography Finish: Includes a dedicated post-processing group (Film Grain, Chromatic Aberration) to eliminate that "AI look" and give you a raw, organic camera feel.🛠️ How to Use:Clothing Reference: Upload any outfit you want to "steal" the look from.Your Influencer: Upload your fixed character (LoRA or Face reference).Run & Generate: Watch as the workflow extracts the clothes and places your influencer in endless, dynamic scenarios (Street style, cafes, studios, etc.).📦 Requirements:Model: Flux.2 Klein (FP8 or FP4 recommended).Control: Ensure you have the required Yolo and Qwen-VL models installed via ComfyUI Manager.Nodes: Simply click "Install Missing Nodes" in ComfyUI Manager.
Flux.2 Klein 9B Workflows (includes Edit workflow + inpaint) + SageAttention + SeedVR2 Upscale
My TG Channel - https://t.me/StefanFalkokAIMy TG Chat - https://t.me/+y4R5JybDZcFjMjFiHi! I introduce my working workflows with Flux.2 Klein 9B for ComfyUI.I have included workflows with Flux.2 Klein 9B modelsYou need to have Flux.2 Klein 9B KV Model (https://huggingface.co/black-forest-labs/FLUX.2-klein-9b-kv/blob/main/flux-2-klein-9b-kv.safetensors), clip (https://huggingface.co/Comfy-Org/vae-text-encorder-for-flux-klein-9b/tree/main/split_files/text_encoders) and vae (https://huggingface.co/Comfy-Org/flux2-dev/tree/main/split_files/vae)if you need Base version - https://huggingface.co/black-forest-labs/FLUX.2-klein-base-9B/blob/main/flux-2-klein-base-9b.safetensorsalso recommend download Flux 2 Klein 9B by wikeeyang, it gives a better result - https://civitai.com/models/2339723/flux2-klein-9b-truehttps://huggingface.co/datasets/StefanFalkok/ComfyUI_portable_torch_2.11.0_cu130_cp313_sageattention_triton - my new ComfyUI build with torch 2.11.0cu130 + sageattention + my workflows, also u need download and install CUDA 13.0 (https://developer.nvidia.com/cuda-13-0-0-download-archive) and VS Code (https://visualstudio.microsoft.com/downloads/)Leave comments if you have trouble or you found the problem with workflows
Pantomime | Facial expression sprite generator
This workflow generates a new facial expression using Flux2.Klein, then refines it with an SDXL model. In the end, you get the full image, and an image of only the face. This could be useful for game creation.I've been thinking about this project for a while, but I've been busy with other things. Initially, I wanted to do it exclusively in SDXL, but I got tired of struggling with just SDXL.
FLUX.2 Klein 9B — Ultimate 6-in-1 Workflow (Face Swap · Inpaint · Auto-Mask · NAG · Refine · Upscale) — 8GB VRAM
FLUX.2 Klein 9B — Ultimate All-in-One ComfyUI Workflow v2.090 nodes · 6 pipelines · 12 organized groups · 8GB VRAM compatibleOne workflow file. Everything you need. No more switching between different JSON files for different tasks.6 Pipelines in 1 WorkflowPipeline: What It Does...🎨 F1: Text → ImageStandard txt2img at 1024×1024🖼️ F2: Single-Reference KV EditEdit specific elements while preserving everything else🚀 F3: Face + Pose SwapTake a face from Image A + pose from Image B → combine realistically🎭 F4: InpaintingManual mask or AI auto-masking → regenerate only masked area🔀 F5: Image MergeBlend two images with adjustable ratio✨ RefinerDetail enhancement, lighting correction, skin texture improvementBuilt-in Features🧭 NAG (Normalized Attention Guidance) — Negative prompts that actually work on distilled Flux🤖 Florence2 Auto-Masking — Type what to mask ("Segment the shirt") → pixel-perfect mask⬆️ 4x UltraSharp Upscaler — Clean upscaling with detail preservation🔷 Tiled VAE Decode — All VAE decodes are tiled to prevent OOM on 8GB VRAM🔗 2-Slot LoRA Chain — Enhancer LoRA always last, add your own LoRAs in the empty slot📋 Built-in Guide Notes — Download links, prompting tips, and troubleshooting right inside the workflowHardware RequirementsMinimum: 8GB VRAM (RTX 3060/4060/4070 with GGUF Q4 model)Recommended: 10-12GB VRAM (RTX 3080/4070 Ti with FP8 model)RAM: 16GB+ComfyUI: Version 0.9.2 or newerModels You NeedDiffusion Model (pick one):FP8: flux-2-klein-9b-fp8.safetensors → DownloadGGUF Q4 (8GB VRAM): → DownloadText Encoder:qwen_3_8b_fp8mixed.safetensors → DownloadVAE:flux2-vae.safetensors → DownloadUpscale Model:4x-UltraSharpV2.pth → DownloadLoRA:klein_9b_enhancer_v2.safetensors → Put in models/loras/downloadCustom Nodes (install via ComfyUI Manager):rgthree-comfy (Image Comparer)ComfyUI-Florence2 by kijai (auto-masking, optional)ComfyUI-GGUF by city96 (only if using GGUF models)How to UseLoad the JSON in ComfyUIuse comphyui manager to install any missing node. ( critical step )Only the Refiner is active by default — everything else is bypassedTo activate a pipeline: right-click its group header → Set Nodes Mode → Always ExecuteTo deactivate: right-click → Set Nodes Mode → Bypass ( or bypass groupe nodes )Read the built-in Note nodes for prompting tips and download linksPrompting TipsFLUX.2 Klein reads prompts like a novelist, NOT like Stable Diffusion.Use: Subject → Setting → Details → Lighting → Technical termsPower words: clean digital file, histogram equalization, gamma correction, studio strobe, shot on Canon EOS R5Avoid: beautiful, masterpiece, best quality, trending on artstationFor text in images: Put text in quotes → A sign reading "HELLO"TroubleshootingOOM: Already using Tiled VAE. Try GGUF Q4 model if still crashingBlack images: Wrong text encoder. 9B needs Qwen3 8B, not 4BPlastic faces: Make sure enhancer LoRA is active (strength 0.7)NAG not working: Install ComfyUI-NAG custom node as alternativeBuilt and tested on RTX 4060 Mobile (8GB VRAM). If you have issues, leave a comment — I'll help you fix it.If this workflow is useful to you, please leave a review — it helps others find it.VERSION HISTORY:v2.1 — Initial public release. +90 nodes, 6 pipelines, NAG, Florence2, Tiled VAE.added groupe bypasser, notes for new people to comphyui
Overtli LLM Studio Suite
A ComfyUI suite of nodes for Pollinations, LM Studio, Copilot CLI, and OpenAI-compatible generation with prompt enhancing, image gen, video gen, speech/audio gen, for local/cloud multi-engine workflows.You can download a sample workflow with it and Flux2Klein 9b GGUF (can always switch model loader).Access the github here for comfyui install:github.com/OvertliDS/overtli-studio-suite## 🚀 Installation & Quick Start ### Installation 1. Navigate to your ComfyUI custom nodes directory and clone this repository: ```powershell cd ComfyUI/custom_nodes/ git clone https://github.com/OvertliDS/overtli-studio-suite.git ``` 2. Navigate into the cloned folder and install the required dependencies using your ComfyUI Python environment: ```powershell cd overtli-studio-suite pip install -r requirements.txt ``` 3. Restart ComfyUI. ### Quick Start Workflow *You can find the nodes by simply searching for "Overtli".* 1. Add the **`GZ_ProviderSettings`** node and save your keys or URLs once. 2. Add the **`GZ_AdvancedTextEnhancer`** or **`GZ_LLMTextEnhancer`** node. 3. Choose your `provider` and `active_engine`. 4. Connect optional `IMAGE` or `AUDIO` inputs when needed. 5. Execute and use the native output directly in your workflow. ### Use With Existing ComfyUI Workflows (CLIP Text Encode) You can use OVERTLI as a **drop-in prompt enhancer** without rebuilding your graph. 1. Add any text-capable OVERTLI node (for example `GZ_TextEnhancer`, `GZ_LLMTextEnhancer`, or `GZ_AdvancedTextEnhancer` with `active_engine = text`). 2. Connect the node's `STRING` output to your existing **CLIP Text Encode** node `text` input. 3. Keep the rest of your workflow unchanged (sampler, model, VAE, etc.). This lets you enhance or rewrite any prompt while still using your current ComfyUI pipeline. ### Example Workflow For users who want a sample workflow with Flux2Klein 9B GGUF, use workflows/OvertliStudioSuite_x_Flux2Klein9B-GGUF.json from this repo. ----- ## ✨ Core Capabilities * **Unified Advanced Routing:** Manage multi-engine workflows through a single, powerful provider node. * **Native Media Contracts:** Full support for standard ComfyUI output types. * `AUDIO` for text-to-speech and text-to-music generation. * `VIDEO` for video generation capabilities. * The Advanced Router dynamically returns `STRING + IMAGE + VIDEO + AUDIO`. * **Native Previews:** Fully integrated with ComfyUI’s native image and text preview rendering. ----- ## 🧩 Registered Nodes | Node | Purpose | Supports | | :--- | :--- | :--- | | ⚡ **`GZ_AdvancedTextEnhancer`** | **(Recommended)** Unified provider/engine router. | Provider-aware engine routing: `text`, `image`, `video`, `text_to_speech`, `speech_to_text`, `text_to_music` with validation. | | 🔌 **`GZ_OpenAICompatibleTextEnhancer`** | Dedicated OpenAI-compatible all-engines node. | `text`, `image_gen`, `video_gen`, `text_to_speech_gen`, `speech_to_text_gen`, `text_to_music_gen` through OpenAI-compatible APIs. | | 📝 **`GZ_TextEnhancer`** | Pollinations text and optional vision enhancement. | Pollinations text generation, optional vision image context, grouped mode presets. | | 🎨 **`GZ_ImageGen`** | Pollinations image generation. | Pollinations image generation models. | | 🎬 **`GZ_VideoGen`** | Pollinations video generation. | Pollinations video generation models (`VIDEO` output). | | 🗣️ **`GZ_TextToSpeech`** | Pollinations speech generation. | Pollinations text-to-speech models (`AUDIO` output). | | 🎧 **`GZ_SpeechToText`** | Pollinations speech transcription. | Pollinations speech-to-text models (`STRING` transcript output). | | 🎵 **`GZ_TextToAudio`** | Pollinations text-to-music generation. | Pollinations text-to-music generation (`AUDIO` output). | | 🖥️ **`GZ_LLMTextEnhancer`**| **(Recommended for local)** Local/OpenAI-compatible text+vision route (LM Studio, Ollama, similar endpoints). | Text generation with optional image context over local/OpenAI-compatible chat endpoints. | | 🤖 **`GZ_CopilotAgent`** | GitHub Copilot CLI route. | Copilot CLI text generation with optional image attachment context. | | ⚙️ **`GZ_ProviderSettings`** | Persisted provider settings helper. | Save/load provider model, endpoint, and API key settings for this suite. | | 📚 **`GZ_PromptLibraryNode`** | Prompt library CRUD/refresh utility. | Prompt management and reusable preset selection. | | 🥞 **`GZ_StyleStackNode`** | Reusable style stack utility. | Composable style bundles for prompt layering across nodes. | ----- ## 🌐 Provider Coverage ### ✨ Pollinations.ai Full support for text, image, video, text-to-speech, speech-to-text, and text-to-music. ### 🖥️ LM Studio Local text generation with optional image context for vision-capable local models. ### 🤖 GitHub Copilot CLI Local CLI-based text enhancement with optional image context. ### 🔌 OpenAI-Compatible APIs Available through both: * `GZ_OpenAICompatibleTextEnhancer` (dedicated node) * `GZ_AdvancedTextEnhancer` with `provider = openai_compatible` Supported engines: * `text` * `image` * `video` * `text_to_speech` * `speech_to_text` * `text_to_music` > **Note:** exact modality availability still depends on the target OpenAI-compatible provider and selected model. The suite now exposes these engines with explicit runtime validation and terminal-visible errors when a provider/model endpoint cannot satisfy a request. ----- ## 🔑 Provider Setup & Authentication The Overtli Studio Suite is designed to be as "zero-config" as possible, supporting cloud, local, and standard API protocols. ### ✨ Pollinations.ai (Cloud) API Key: Visit enter.pollinations.ai to generate your key. Setup: Enter your key into the GZ_ProviderSettings node or set the GZ_POLLINATIONS_API_KEY environment variable, as well as support in Pollinations nodes to enter api key directly. Pricing: `Free` tagged models allow free generations, but pollen usage is determined based on your free plan tier within enter.pollinations.ai so aim to reach higher tiers for more free generations or top up using gems while waiting for your pollen to reset. `Paid` models are supported as well if gems have been purchased. ### 🤖 GitHub Copilot CLI (Cloud/Local Hybrid) Auto-Detection: If you are already signed in via the GitHub CLI (gh auth login), Overtli will automatically detect your session and "just work." Prerequisites: Requires an active Copilot subscription and the GitHub CLI installed (It will open a copilot-cli terminal silently by default for communication). Vision/Image context: OVERTLI writes Comfy `IMAGE` inputs to temporary local files and references them with Copilot CLI `@path` attachment syntax so vision-capable models receive the actual image bytes. Background behavior: Copilot auth recovery and vision-cache retry handling are managed automatically in the background so the node surface stays simple. With `vision_enabled` on, prior runtime vision-cache entries do not silently block new image attempts by default. ### 🖥️ LM Studio (Local) Host: Defaults to http://localhost:1234. Setup: Ensure LM Studio is running and the "Local Server" is started. No API key is required by default. Vision: Support for vision-capable local models is integrated directly into the GZ_LLMTextEnhancer. ### 🔌 OpenAI-Compatible APIs (Custom) Flexibility: Use any provider that supports the OpenAI standard (e.g., Groq, Together AI, LocalAI). Setup: Requires a Base URL and API Key configured in the GZ_ProviderSettings node (or just use the GZ_AdvancedTextEnhancer or GZ_LLMTextEnhancer node). ----- ## 🧠 Prompting & Output Contracts ### The Composition Model All primary routes follow the same strict sequence: 1. **Custom Instructions** 2. **Selected Mode Preset** 3. **Raw Prompt** (User Prompt) 4. **Style Layers** (`style_preset_1..3` + `additional_styles`) > 💡 **Tip:** The `additional_styles` input is socket-first (`forceInput`) and is designed to chain seamlessly from the `GZ_StyleStackNode`. ### Quick Instruction Modes Overview Instruction presets are grouped by task family and can be toggled on/off per route. At-a-glance preset counts (current build): - **Total instruction presets:** **63** (excluding `Off` options) - **Text:** 14 - **Image:** 24 (4 vision + 10 generation + 6 editing + 4 narrative) - **Video:** 11 (8 generation + 3 analysis) - **TTS:** 8 - **Text-to-Audio:** 3 - **Speech-to-Text:** 3 - **Text modes:** examples include `Enhance` (expand a short idea into a richer prompt), `Refine` (tighten wording while preserving intent), and `Translation Prompt` (convert to generation-ready English). - **Image modes:** examples include `Tags` (comma-separated visual tags), `Simple Description` (single concise visual description), `Concept Art` (design-forward scene direction), and `Upscale Image Prompt` (detail-preserving enhancement guidance). - **Video modes:** examples include `Cinematic Video Prompt` (film-style motion direction), `Loop Video Prompt` (seamless repeating motion guidance), and `Video Summary` (chronological visual recap of footage). - **TTS modes:** examples include `Voiceover Script` (clean narration pacing for speech synthesis), `Character Voice` (expressive dialogue delivery), and `SSML Enhancement` (markup-driven timing and emphasis control). - **Text-to-Audio modes:** examples include `Audio Prompt Enhance` (clarify generation intent), `Music Direction` (genre/instrument/energy guidance), and `Ambience and Foley` (layered environmental sound design). - **Speech-to-Text modes:** examples include `Clean Transcript` (readability cleanup while preserving meaning), `Punctuation and Casing` (restore sentence structure), and `Structured Notes` (concise key points and actions). If a mode is set to `Off`, the route relies on custom instructions + raw prompt + style layers. *Only enable ONE instruction group* ### Quick Style Layering Overview - **Built-in style library:** **656** presets (+ `Off` option) - **Built-in style slots:** up to **3** per primary node (`style_preset_1..3`). - **Extended style stack:** connect `additional_styles` from **`GZ_StyleStackNode`** for up to **7** extra styles. - **Total style layers available:** up to **10** (3 built-in + 7 stacked). Style examples (brief): - **Photograph (Real Life):** grounded, natural realism with real lens/lighting behavior. - **Cinematic Still:** film-like framing, layered depth, and motivated key/fill/rim lighting. - **Anime Illustration:** expressive line work, controlled cel-shading, and clean silhouette readability. - **Line Art (Clean):** minimal contour-focused visuals with crisp outlines and low noise. - **Concept Art:** production-style world/design visualization with clear value hierarchy. ### Output Contracts | Route Type | Native ComfyUI Output | | :--- | :--- | | **Text** | `STRING` | | **Image** | `IMAGE` | | **Video** | `VIDEO` | | **Audio-producing** | `AUDIO` | | **Advanced Router** | `STRING`, `IMAGE`, `VIDEO`, `AUDIO` | ----- ## ⚙️ Settings and Precedence Persistent settings are stored locally in: ```text ComfyUI/user/overtli_studio_settings.json ``` **Resolution Precedence:** 1. Runtime node input *(Highest priority)* 2. Environment variable 3. Persisted settings 4. Suite default *(Lowest priority)* <details> <summary><b>🔧 Click to view useful Environment Variables</b></summary> * `GZ_POLLINATIONS_API_KEY` * `GZ_POLLINATIONS_CHAT_TIMEOUT` * `GZ_POLLINATIONS_IMAGE_TIMEOUT` * `GZ_POLLINATIONS_VIDEO_TIMEOUT` * `GZ_POLLINATIONS_TTS_TIMEOUT` * `GZ_POLLINATIONS_STT_TIMEOUT` * `GZ_POLLINATIONS_AUDIO_TIMEOUT` * `GZ_LMSTUDIO_HOST` * `GZ_LMSTUDIO_PORT` * `GZ_LMSTUDIO_API_KEY` * `GZ_COPILOT_EXECUTABLE` * `GZ_COPILOT_MODEL` * `GZ_COPILOT_TIMEOUT` * `GZ_OPENAI_COMPAT_BASE_URL` * `GZ_OPENAI_COMPAT_API_KEY` * `GZ_OPENAI_COMPAT_MODEL` * `GZ_LOG_LEVEL` * `GZ_TEMP_DIR` </details> ----- ## 🧯 Troubleshooting **OpenAI-compatible route does not appear as its own node** That is expected. Use `GZ_AdvancedTextEnhancer` with `provider = openai_compatible`. **STT fails with missing input** `GZ_SpeechToText` requires an `AUDIO` input connection to function. **Video generation blocks the UI** That is expected for synchronous provider-side video generation. The UI will resume once rendering is complete. **LM Studio/Ollama does not return models** Make sure LM Studio/Ollama is running and exposing its API endpoint correctly. **Copilot route fails unexpectedly** Check that: * GitHub Copilot CLI is installed and reachable. * The configured executable path is correct if adjusted manually. * Your authentication session is valid. * Environment token overrides are not conflicting. ----- ## 🤝 Support & Funding If this suite accelerates your workflow, consider supporting the continued development of Overtli tools and other projects using the GitHub Sponsor link on the repository sidebar. ## 📄 License MIT. See `LICENSE`.
Consistent Character Dataset Creator (Klein_9b)
I started creating Loras locally, so needed a dataset creator.Uses Klein 9_B with any good facial image to create 20 different images.也有中文说明Operates fairly well on Low VRAMrun 2 or 3 times on a random seed and select the good ones. Instagram: https://www.instagram.com/synth.studio.models/Buy me a☕ https://ko-fi.com/lonecatoneThis represents many of hours of work. If you enjoy it, please 👍like, 💬 comment , and feel free to ⚡tip 😉
Klein_9b Wildcard Workflow
Note: If you have ANY issues with nodes not downloading, read the notes or reach out. There's nothing that special about any of them that aren't core modules.⛔⚠️🛑✋ Read the notes completly before using. Most common install and node problems are listed in the directionsBy numerous requests, This is a more advanced version of my Klein_4b workflow也有中文说明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-breakdownWildcard workflow:Designed to be used with or without character LorasRandom Prompt generation or use your ownInstructions on how to easily create yor own wildcards and random promptsRandom Lora Loader stacksSeperate Lora loader for CharacterI2I for character facesSeed VR2 Upscaler to get 4k on low VRAMEssential Post Production Suite也有中文说明Instagram: https://www.instagram.com/synth.studio.models/Buy me a☕ https://ko-fi.com/lonecatoneThis represents countless of hours of work. If you enjoy it, please 👍like, 💬 comment , and feel free to ⚡tip 😉
Flux 2D & Klein_9b (High & Low VRAM), w/ Controlnet, GGUF capable
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-breakdownEach Flux model has it's own combination of text encoders. Read the notes carefully. You will get MATT errors of it is incorrect.也有中文说明Ver 5.0Added controlnetAdded a sentence remover to prevent CivitAi FlagsVer 4.0Total overhaulOptimized for Klein_9bOptions to control the schedulerUpgraded detailersUpgraded Post Production Suite📷 Optical Realism upgradeArcane Bloom FXLensFXClarityFX也有中文说明Flux 2D I2I, T2I, Upscaler, Detailers, and a hefty Post Production SuiteOptimized for Klein as well as Flux 2DInclude the Klein Enhancer (Similar to a detail daemon) to improve the quality of the Quant imagesPrompt AssistAllows up to 10 images to be loaded for referenceUltimate Upscaler packageSeed VR2 UpscalerFace, Hand, Eye, and Expression DetailersAdd lighting, color correction , and specicl effects after generation in the Post Production SuiteFlux 2D Base ModelFlux 2D Model (You have to agree to their terms first):https://huggingface.co/black-forest-labs/FLUX.2-dev/tree/mainFlux 2D GGUF Model https://huggingface.co/gguf-org/flux2-dev-gguf/tree/mainFlux 2D Mistral text endoders https://huggingface.co/Comfy-Org/flux2-dev/tree/main/split_files/text_encodersFlux 2Dev Turbo Lora https://civitai.com/models/2266313/flux2-dev-turbo-lora?modelVersionId=2550958Flux 2 Klein Model & Text encoderKlein 4bText Encoder qwen_3_4b.safetensorsFlux 2 Klein diffusion_model Flux-2-klein-base-4b-fp8.safetensorsKlein 9bText Encoder Qwen_3_8b.safetensorsFlux 2 Klein 9b FLUX.2-klein-9BVAE (for both)VAE flux2-vae.safetensorsInstagram: 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 😉
My Klein Workflow
This is a simple workflow that takes from my old One-Shot workflow. way less needlessly complex while doing pretty much the same thing. I apply a lut and skin upscaler towards the end.this is the skin upscaler you will need https://openmodeldb.info/models/1x-SkinContrast-High-SuperUltraCompact. i was unable to find the specific LUT i use, but it's more of a personal preference thing anyway, you can find a bunch here https://freshluts.com/.
Flux.2 Klein + Refiner
Flux.2 Klein + Refiner — Clean, Fast, No NonsenseI'll be honest: I was almost ready to give up on Flux.2 Klein.I spent an entire night downloading and testing workflows I found online — some popular, some less so — and none of them worked the way I needed. Generation times were absurd, sometimes hitting 400 seconds or more. On top of that, I kept getting images at resolutions I never asked for, with no upscale nodes in sight and no clear explanation. I was going crazy trying to fix issues I didn't even fully understand.At some point I thought: maybe this is just how Flux.2 Klein works. But something didn't feel right. So instead of giving up, I built my own workflow from scratch — adapting it to the way I actually work.Turns out the answer was simpler than I expected.What this workflow doesThis is a clean, straightforward Flux.2 Klein workflow with an optional SDXL refiner pass. No bloated node chains, no mystery outputs, no hidden upscaling. What you set is what you get.Generation times I've achieved:Without refiner: 20–30 seconds per generationWith refiner (SDXL 1.0 DMD): 35–40 seconds(obviously based on my hardware, so your results may vary)Resolution: fully customizable via a dedicated node — I personally use 1536×1920, but you can set whatever you need.How the refiner worksAfter Flux.2 Klein generates the base image, a second KSampler with an SDXL 1.0 DMD checkpoint refines it at low denoise. Nothing exotic — just a classic refiner setup that adds detail and polish without rebuilding the image from scratch. You can run it or skip it depending on your needs.Why am I sharing this?Because I wasted hours on workflows that looked impressive but performed terribly, and I don't want the same to happen to you. This one works (at least for me :)... ). It's not fancy, but it's fast, predictable, and the results speak for themselves.If you try it, let me know what you think. And if you use it for portraits or lifestyle shots, I'd love to see what you create.All sample images were generated with this workflow. No post-processing.
ComfyUI ALL-in-One Fast Flux.2 Klein Raster to Vector - Image to Image (With Prompt Saver) Workflow
This is a very simple, beginner-friendly, fast ComfyUI workflow based on Flux.2 Klein model (4B or 9B) that can first generate an useual Raster Image file (.jpg or .png or .webp) image-to-image output then right after that it converts it again to Vector Image file (.svg) output on the fly. This workflow works great for illustration-style images, like stickers and cartoons. This workflow is built upon my previously published Flux.2 Klein Text-To-SVG Workflow that you can find in my CivitAI Profile.This workflow uses a LORA that I trained extensively on Flux.2 Klein (I have two versions, one for 4B model and another for 9B model) with 250 high resolution, crisp & clear, meticulously selected digital artworks of multiple varieties so that the end results can be as fine as possible (See the Required Models section for the link to get the correct version of the LORA). Normally Flux.2 Klein has a very strong bias for AI Digital Photgraphy style outputs or near photorealistic outputs, but my LORA takes advantage of Flux.2 Klein's robust output generation speed but guides it forward to focus more on digital arts and simple vector illustrations.I have implemented my own Prompt Saver Subgraph here so it can save Text to Image Generation Data into a human readable .txt file. This will automatically get and write your metadata to the .txt file. This workflow also uses Flux.2 Klein Enhancer for quality outputs. You will find all the saved prompt files that it generated with the images (.jpeg and .svg) inside the Archive (.Zip) that has the workflow. Also with the Image Saver Simple node used you may embed the workflow itself with each saved image or save the image and workflow for your work separately.Make sure that you have latest enough versions of both ComfyUI and ComfyUI manager to manage and install any missing dependencies (missing nodes, patches etc.) to use this workflow properly.#### Very Very Important :Even before loading this workflow into ComfyUI and install nodes needed using ComfyUI Manager you must go to your ComfyUI's python environment and run this command to install necessary python packages to handle Raster Images (.jpeg or .png or .webp) to Vector Images (.svg) conversion - python3 -m pip install blend_modes vtracer PyWaveletsRequired Files==================Flux.2 Klein 9B Models ( If you are using 6GB or 8GB VRAM GPU )-+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++### Download Link for Flux.2 Klein 9B GGUF Models--------------------------------------------------Only for this workflow I used Flux.2 Klein 9B Q2 GGUF Model (lowest Quant for 9B), but usually I useGet the Q2 Model from here (One used in this workflow that helped me reduce system power usage) -https://huggingface.co/unsloth/FLUX.2-klein-9B-GGUF/resolve/main/flux-2-klein-9b-Q2_K.ggufGet the Q4 Model from here (One used in my all other Flux.2 Klein 9B workflows, it offers better performance & results ) -https://huggingface.co/unsloth/FLUX.2-klein-9B-GGUF/resolve/main/flux-2-klein-9b-Q4_K_M.gguf### Download Link for Flux.2 Klein 9B Text Encoder---------------------------------------------------https://huggingface.co/Comfy-Org/flux2-klein-9B/resolve/main/split_files/text_encoders/qwen_3_8b_fp8mixed.safetensorsFlux.2 Klein Common VAE Model -++++++++++++++++++++++++++++++### Download Link for Flux.2 Klein VAE (Works on both 4B and 9B)------------------------------------------------------------------https://huggingface.co/Comfy-Org/flux2-dev/resolve/main/split_files/vae/flux2-vae.safetensorsAlso consider using these 4B Model files (even GGUF Text Encoders) if you are using on a significantly weaker GPU ( with only 2GB or 4GB VRAM ) ---Flux.2 Klein 4B Models -+++++++++++++++++++++++++### Download Link for Flux.2 Klein 4B GGUF Model-------------------------------------------------This is for 4GB VRAM -https://huggingface.co/unsloth/FLUX.2-klein-4B-GGUF/resolve/main/flux-2-klein-4b-Q4_K_S.ggufThis is for 2GB VRAM -https://huggingface.co/unsloth/FLUX.2-klein-4B-GGUF/resolve/main/flux-2-klein-4b-Q2_K.gguf### Download Link for Flux.2 Klein 4B Text Encoder---------------------------------------------------I used this one on my Flux.2 Klein 4B workflows, it may work for you for 4B model. Note that 4B text encoder will not work well with 9B models and 9B text encoder will not work well with 4B models and on some low end systems you may even have to find out GGUF text encoder replacement for this one -https://huggingface.co/Comfy-Org/flux2-klein-4B/resolve/main/split_files/text_encoders/qwen_3_4b.safetensorsLORA Used on this workflow -+++++++++++++++++++++++++++I trained my own LORAs (for Flux.2 Klein 9B & 4B) to effectively produce quality vector drawing, artwork, silhouettes, stencils, minimal drawings, logos & other minimal vector digital AI arts. You can download the one that will work with your selected Flux.2 Klein model (9B or 4B) suited for your needs and system's configuration. Both are hosted on CivitAI. Both can perform without any reference image at all (for Txt2Img) or with a single Raster reference image (for Img2img).Simple Fine Vector Flux.2 Klein 9B-----------------------------------https://civitai.com/models/2462137?modelVersionId=2768352Simple Fine Vector Flux.2 Klein 4B-----------------------------------https://civitai.com/models/2462142?modelVersionId=2768357
flux2 klein - Steal Tattoo / Body Writing
Steal someones tattoo! Or body writing!Download the workflow and necessary lorashttps://civitai.com/models/2506194/place-tattoohttps://civitai.com/models/2506180/flux2-klein-extract-tattooUse with the prompt "place this tattoo ..."Examples:place this tattoo on her chinplace this tattoo on her backplace this tattoo on her armUse the same type of prompting even if you're doing body writing
Smart Text Styler Pro (Node for Flux2 Klein)
Smart Text Styler Pro (Node for Flux2klein)📄 DescriptionSmart Text Styler Pro is a powerful custom node for ComfyUI, specifically designed to create highly stylized text and perfect typography for Flux2klein.The node provides two synchronized outputs:Layout Image (Mask): A clean text layout on a solid background, respecting precise positioning and professional Action Safe Areas.Flux Optimized Prompt (String): An advanced instructional prompt that guides the AI to "texture" the letters while strictly preserving their shape and placement.🚀 Key FeaturesMulti-line Support: Individual control over font size and letter spacing for every single line.Layered Style System: A unique stacking logic: Material → Decoration → Action → Environment. Create complex combinations like "Rusty Metal + Bullet Holes + Burning + Sparks".Triple Environment Slots: Add up to three atmospheric effects simultaneously (rain, lightning, smoke, glow).Smart Background System: Choose from built-in presets (Underwater, Space, Cyberpunk City), use custom prompts, or a simple solid color.Scene Interaction: A dedicated field to describe how the text physically interacts with the environment (e.g., "half-buried in sand", "floating in zero gravity").Action Safe Area: Automatic 10% padding from the edges to ensure your text looks professional and never gets cropped.⚙️ Settings Explained📝 Text & Typographytext_lines: Enter your text here. Use Enter for new lines.font: Choose any font from your /fonts folder (supports .ttf and .otf).font_sizes: Comma-separated sizes (e.g., 120, 60). If there are more lines than values, the last value will repeat.letter_spacing: Comma-separated spacing (e.g., 0, 5). Use this to stretch or compress specific lines.line_spacing: Vertical distance between lines in pixels.🎨 Canvas & Positioningwidth / height: The output resolution of the layout image.position: Alignment (Center, Top, Bottom-Left, etc.). Note: The node automatically applies an Action Safe Area (~10%) to prevent text from touching the edges.canvas_color: Base color (Black, White, Transparent, etc.). Black is highly recommended for Flux workflows.🧱 Style Core (Prompt Building Logic)material: The base substance (Wood, Stone, Metal, Lava, Chocolate, Neon, etc.).decoration: Surface details (Vines, Bullet Holes, Rust, Cracks, Tattoos).action: Dynamic state (Burning, Melting, Exploding, Frozen).environment_1, 2, 3: Atmospheric effects surrounding the text (Sparks, Smoke, Lightning, Snow).🖼️ Background & Interactionbg_mode:simple: Uses the canvas_color.preset: Choose from curated scenes (Underwater Corals, City Night Neon, Space Nebula, etc.).custom: Enter your own background description.scene_interaction: Crucial Field. Describes physical contact with the world.Examples: "standing on the sandy bottom", "floating in zero gravity".⚡ Advanced Settingsrandom_style: Enable to let the node randomly pick materials, actions, and effects.extra_instruction: Add your own specific tokens to the final prompt.style_mode: Lighting presets (Realistic, Creative, 3D Render).intensity: Level of detail (adds "highly detailed" modifiers). 📥 InstallationDownload the ComfyUI-SmartTextStyler folder and place it in ComfyUI/custom_nodes/.Manage Fonts:Navigate to ComfyUI/custom_nodes/ComfyUI-SmartTextStyler/fonts.Cleanup: Delete any default fonts you don't need to keep your UI menu clean.Add Your Own: Drop your favorite .ttf or .otf files into this folder.Tip: High-contrast, bold, and display fonts work best for AI texturing!Restart ComfyUI.📝 Example "Scene Interaction" PhrasesTry these in the scene_interaction field for stunning results:Underwater: Text resting on the sandy bottom, surrounded by tropical fishSpace: Text floating in zero gravity, faint star reflections on the surfaceCyberpunk: Text standing on wet asphalt, neon reflections in rain puddlesForest: Text half-buried in thick green moss, ancient roots wrapping around the lettersDesert: Text half-buried in wind-swept sand dunes, heavy erosion detailsThe archive contains a node folder and a simple workflow.
Flux2 Easy Swap: Head, Face or Outfit (No Lora)
🚀 Flux2 Easy Swap: Head, Face or OutfitAbout V1.1Fixing Minor ıssues. Please upgrade your photobuf and mediapipe for "personmask" ( python_embeded\python.exe -m pip install --upgrade mediapipe protobuf google-cloud-videointelligence )This workflow is optimized for the Flux model to perform natural and high-quality swaps of faces, heads, or outfits using PersonMaskUltra V2.***HEAD OR FACE SWAP***In the PersonMask NodeImg1 enable; body, clothe,accecories backgroundImg2 enable; face, hairPrompt; It's best to leave this blank. Don't write anything. When you write something, the head size appears large and the colors may not match.orPrompt; "Replace the entire face and hair-style in image2 with the person image1. Make the face and body proportions perfect in the final image."***OUTFIT CLOTHING SWAP***In the PersonMask NodeImg1 enable; face,hair,body,backgroundImg2 enable; clothesPrompt; "Style the person in image1, with every article of clothing in image2. In the final image, all the features of the clothes should be on the person in image1."☕ Support My Work👉 https://buymeacoffee.com/ugurdoyduk
FLUX.2-klein-Designated character face swap
Video tutorial：Usage method：1、upload the original image2、upload face reference reference map3、fill in the character position you want to replace on the imageWe recommend trying it out online before downloading; it's completely free!Online experience address:https://www.runninghub.ai/post/2036670076690112513/?inviteCode=rh-v1058
FLUX.2-klein-Text-to-Image-Image Editing
Video tutorial：Usage instructions:*Image effect switch on means using image editing, fully off means using text-to-image）1、Fill in the width and height (only needed when using text-to-image）2、Fill in the prompt3、Whether to use local redraw (only for image 1), after turning it on, apply a mask to image 1 (there is a button in the lower right corner of the image to select the mask range)4. Upload the corresponding image and turn on the switch as needed, use in order 1, 2, 3 (do not delete the image even if it is not used, otherwise an error will occur)We recommend trying it out online before downloading; it's completely free!Online experience address:https://www.runninghub.ai/post/2034521906816225282/?inviteCode=rh-v1058
FLUX.2-klein-Headswap-Single Version
Video tutorial：Usage method：1、upload the original image2、upload face reference reference mapWe recommend trying it out online before downloading; it's completely free!Online experience address:https://www.runninghub.ai/post/2036297825587306498/?inviteCode=rh-v1058
FLUX.2-klein-Designated character face swap
Video tutorial：Usage method：1、upload the original image2、upload face reference reference map3、fill in the character position you want to replace on the imageWe recommend trying it out online before downloading; it's completely free!Online experience address：https://www.runninghub.ai/post/2036670076690112513/?inviteCode=rh-v1058
nnegret | klein three-view drawing workflows | 三视图生成工作流
Q群：173273201💻 Requirements / 硬件要求VRAM / 显存: 16GB minimum for a smooth experience.建议显存 16G 及以上以确保流畅运行。Environment: Latest ComfyUI version recommended.推荐环境：ComfyUI 最新版。🛠️ How to Use / 使用指南Update: Ensure your ComfyUI is up to date.更新：确保你的 ComfyUI 处于最新版本。Missing Nodes: Use ComfyUI-Manager -> Install Missing Custom Nodes.补齐节点：使用 Manager 插件点击“安装缺失节点”即可自动补齐。Model Path: Place Flux2Klein models in the correct checkpoints or unet folders.模型路径：确保模型已放置在对应的路径下。Drag & Drop: Drag the JSON file into ComfyUI and start creating!开始使用：直接将 JSON 文件拖入即可开始创作。✉️ Author's Note / 作者的话[EN] I know how overwhelming a "spaghetti" of nodes can be. I hope this clean starting point helps you focus on creativity rather than troubleshooting. If you like it, please leave a star or share your images![CN] 我深知新手面对“乱如麻”的节点时的困惑。希望这个清爽的起点能让你专注于创作，而非排查错误。如果你觉得好用，请给我一个五星好评或分享你的返图！
Flux2 Klein Inpainting + Image Edit Simple Workflow
I wanted a workflow which could do both inpainting and image edit in one but also remains simple.
ComfyUI beginner friendly Flux.2 Klein 9B KV GGUF Simple Flawless Fast Dress Photoshoot Workflow by Sarcastic TOFU
This is a very simple ComfyUI beginner friendly Flux.2 Klein 9B KV GGUF Simple Flawless Fast Dress Photoshoot Workflow that will work with very simple text editing instructions in natural language to generate your desired Dress Photoshoot out four reference images - Picture 1 is a Face Reference, Picture 2 is the Dress ( & maybe Footwear Reference ) with no head visible, Picture 3 is the background of the photoshoot and lastly Picture 4 is the pose reference. The resulting output will closely (but not exactly) match elements from those 4 inputs. See the Prompt_sets.txt file included in the zip file containing the workflow for a general guideline on how to generate the initial inputs if you don't have any inputs and how to combine them in a single output for this workflow. I also included a sample dataset of five images (4 inputs + 1 output) with this for you to have better understanding of this workflow. This uses NO LORA or ControlNet ( officially Flux.2 Klein have not yet provided any viable solution for it but still it can do lots of cool stuffs ).I used a Q2 Flux.2 Klein 9B KV GGUF Model that works well with my 8GB VRAM AMD GPU or even GPUs with lower 6GB VRAM. But is you have better GPU you can simply swap the model with full unquantized Flux.2 Klein 9B KV model or on the other side if you have a much weaker 4GB VRAM or even 2GB VRAM GPU you can find other low grade hardware supported GGUF models from here ( Unsloth's HuggingFace repo for Flux.2 Klein 4B GGUF models - https://huggingface.co/unsloth/FLUX.2-klein-4B-GGUF/tree/main ) and other mathing resources. As this workflow does not use any LORA it can be easily modified to work with any regular or GGUF Flux.2 Klein 9B or 4B model file but you may or not get better results.The FLUX.2 klein 9B KV is an optimized variant of the standard 9B model specifically designed for multi-reference image editing. While the regular 9B model is a powerhouse for high-quality, general-purpose generation, the KV (Key-Value) version is a specialized tool for tasks where you are using the same reference images across multiple generations or edits.This workflow also helps you to save your AI Photoshoot output generation Data into a human readable .txt file. This will automatically get and write your metadata to the .txt file. Also with the Image Saver Simple node used you may embed the workflow itself with each saved image or save the image and workflow for your work separately. In this way a readable .txt file for each run of this workflow will be generated (matching Automatic1111 / EasyDiffusion's .txt outputs).You need a Hugging Face account to download your necessary files (Details are mentioned below). Make sure you install GGUF addon for ComfyUI using ComfyUI manager or any other missing nodes you may have and place the correct files in correct places. Also check out my other workflows for SD 1.5 + SDXL 1.0, WAN 2.1, WAN 2.2, MagicWAN Image v2, QWEN, HunyuanImage-2.1, HiDream, KREA, Kontext, Chroma, AuraFlow, Z-Image Turbo, Flux.1 Dev, Flux.2 Klein 4B and 9B. Feel free to toss some yellow Buzz on stuffs you like.How to use this -#1. Just select your desired GGUF (or swap it with full model) model first and#2. now select your images - Picture 1 for Face, Picture 2 for Dress, Picture 3 for Background and Picture 4 for Pose.#3. on next step enter your image editing instructions.. (see the Prompt_sets.txt file included for a general guideline)#4. then select how many images you want (Change the number besides the "Run" button)#5. after this select image sampling methods, CFG, steps etc. settings (you may wanna stay with defaults)#5. finally press the run button to generate. That's it....The resulting Dress Photoshoot output will very closely (but not exactly) match elements from those 4 inputs. Try to be very precise with your prompt for better results, sometimes even a minor tweak can get you excellent or bad results.Required Files==================Flux.2 Klein 9B Models ------------------------### Download Link for Flux.2 Klein 9B KV GGUF Model used---------------------------------------------------------https://huggingface.co/QuantStack/FLUX.2-Klein-9B-KV-GGUF/resolve/main/Flux-2-Klein-9B-KV-Q2_K.gguf### Download Link for Flux.2 Klein 9B Text Encoder used--------------------------------------------------------https://huggingface.co/Comfy-Org/flux2-klein-9B/resolve/main/split_files/text_encoders/qwen_3_8b_fp8mixed.safetensors### Download Link for Flux.2 Klein VAE used--------------------------------------------https://huggingface.co/Comfy-Org/flux2-dev/resolve/main/split_files/vae/flux2-vae.safetensors
Klein Upscale
Klein 4MP Upscale WorkflowUpscales with Klein using the image both as reference and as latent.Key ParametersParameter Range Effect denoise 0.6 – 0.8 Higher = more change, lower = closer to original reference/prompt 0.3 – 0.8 Lower = more reference influence, higher = more prompt influence s_noise 1.0 – 1.2 Higher = more detail added (1.2 can introduce artifacts)Notesreference/prompt and denoise work together — adjust both in tandemLower reference/prompt (0.3) = strong structure lock, pair with lower denoise (0.6) for tight fidelityHigher reference/prompt (0.8) = reference has less grip, more prompt-driven, pair with higher denoise (0.8) for more creative output
Comfyui-flux2klein-Lora-loader with Auto strength
Architecture-aware LoRA loading for FLUX.2 Klein (9B) in ComfyUI, with automatic per-layer strength calibration based on forensic weight analysis.Same job as my other custom node here : https://civitai.com/models/2436306/comfyui-zit-lora-loader-with-auto-strength but this one is dedicated for Flux2klein 9BFor optimal results I recommend using the "FLux2Klein-Enhancer" : https://github.com/capitan01R/ComfyUI-Flux2Klein-EnhancerIf you find this helpful :) https://buymeacoffee.com/capitan01r
Flux 2 Klein Workflow - t2i / i2i
Flux 2 Klein Workflow :Multiple LoRA switch + t2i or i2i + Upscaler for i2i + Final Upscaler
Mrpopo's Flux.2 Klein 9B anime workflow
unlock the capability of anime and unleash flux klein 9b true potential with my anime workflow. It features color adjustments and a simple upscale, perfect for anime. I've included notes in the workflow to help you out. I strongly suggest using my anime lora with this workflow. I appreciate anyone who tries my models! Cant wait to see what creative images you bring.In the save image node, please change the hash to the new hash from this page. you can see it on the right side of this page. I've included notes to help you with that. Its to upload image to the correct workflow page. Use the AUTOV3 hash. I appreciate it!Also, you can change the AUTOV3 hash of my version 1 anime lora to version 2 (i've updated the lora) in the image saver node.This workflow is also good for non anime generations. just disable the anime lora, choose your sampler and produce photorealistic images. Sometimes simple workflow equals better. no complications. You could increase the resolution as well to 1664x2432. My Lora can handle that high of resolution.
ComfyUI ALL-in-One Fast Flux.2 Klein Raster to Vector Image (With Prompt Saver) Workflow
This is a very simple, beginner-friendly, fast ComfyUI workflow based on Flux.2 Klein model (4B or 9B) that can first generate an useual Raster Image file (.jpg or .png or .webp) text-to-image output then right after that it converts it again to Vector Image file (.svg) output on the fly. This workflow works great for illustration-style images, like stickers and cartoons.This workflow uses a LORA that I trained extensively on Flux.2 Klein (I have two versions, one for 4B model and another for 9B model) with 250 high resolution, crisp & clear, meticulously selected digital artworks of multiple varieties so that the end results can be as fine as possible (See the Required Models section for the link to get the correct version of the LORA). Normally Flux.2 Klein has a very strong bias for AI Digital Photgraphy style outputs or near photorealistic outputs, but my LORA takes advantage of Flux.2 Klein's robust output generation speed but guides it forward to focus more on digital arts and simple vector illustrations.I have implemented my own Prompt Saver Subgraph here so it can save Text to Image Generation Data into a human readable .txt file. This will automatically get and write your metadata to the .txt file. This workflow also uses Flux.2 Klein Enhancer for quality outputs. You will find all the saved prompt files that it generated with the images (.jpeg and .svg) inside the Archive (.Zip) that has the workflow. Also with the Image Saver Simple node used you may embed the workflow itself with each saved image or save the image and workflow for your work separately.Make sure that you have latest enough versions of both ComfyUI and ComfyUI manager to manage and install any missing dependencies (missing nodes, patches etc.) to use this workflow properly.#### Very Very Important :Even before loading this workflow into ComfyUI and install nodes needed using ComfyUI Manager you must go to your ComfyUI's python environment and run this command to install necessary python packages to handle Raster Images (.jpeg or .png or .webp) to Vector Images (.svg) conversion - python3 -m pip install blend_modes vtracer PyWaveletsRequired Files==================Flux.2 Klein 9B Models ( If you are using 6GB or 8GB VRAM GPU )-+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++### Download Link for Flux.2 Klein 9B GGUF Models--------------------------------------------------Only for this workflow I used Flux.2 Klein 9B Q2 GGUF Model (lowest Quant for 9B), but usually I useGet the Q2 Model from here (One used in this workflow that helped me reduce system power usage) -https://huggingface.co/unsloth/FLUX.2-klein-9B-GGUF/resolve/main/flux-2-klein-9b-Q2_K.ggufGet the Q4 Model from here (One used in my all other Flux.2 Klein 9B workflows, it offers better performance & results ) -https://huggingface.co/unsloth/FLUX.2-klein-9B-GGUF/resolve/main/flux-2-klein-9b-Q4_K_M.gguf### Download Link for Flux.2 Klein 9B Text Encoder---------------------------------------------------https://huggingface.co/Comfy-Org/flux2-klein-9B/resolve/main/split_files/text_encoders/qwen_3_8b_fp8mixed.safetensorsFlux.2 Klein Common VAE Model -++++++++++++++++++++++++++++++### Download Link for Flux.2 Klein VAE (Works on both 4B and 9B)------------------------------------------------------------------https://huggingface.co/Comfy-Org/flux2-dev/resolve/main/split_files/vae/flux2-vae.safetensorsAlso consider using these 4B Model files (even GGUF Text Encoders) if you are using on a significantly weaker GPU ( with only 2GB or 4GB VRAM ) ---Flux.2 Klein 4B Models -+++++++++++++++++++++++++### Download Link for Flux.2 Klein 4B GGUF Model-------------------------------------------------This is for 4GB VRAM -https://huggingface.co/unsloth/FLUX.2-klein-4B-GGUF/resolve/main/flux-2-klein-4b-Q4_K_S.ggufThis is for 2GB VRAM -https://huggingface.co/unsloth/FLUX.2-klein-4B-GGUF/resolve/main/flux-2-klein-4b-Q2_K.gguf### Download Link for Flux.2 Klein 4B Text Encoder---------------------------------------------------I used this one on my Flux.2 Klein 4B workflows, it may work for you for 4B model. Note that 4B text encoder will not work well with 9B models and 9B text encoder will not work well with 4B models and on some low end systems you may even have to find out GGUF text encoder replacement for this one -https://huggingface.co/Comfy-Org/flux2-klein-4B/resolve/main/split_files/text_encoders/qwen_3_4b.safetensorsLORA Used on this workflow -+++++++++++++++++++++++++++I trained my own LORAs (for Flux.2 Klein 9B & 4B) to effectively produce quality vector drawing, artwork, silhouettes, stencils, minimal drawings, logos & other minimal vector digital AI arts. You can download the one that will work with your selected Flux.2 Klein model (9B or 4B) suited for your needs and system's configuration. Both are hosted on CivitAI.Simple Fine Vector Flux.2 Klein 9B-----------------------------------https://civitai.com/models/2462137?modelVersionId=2768352Simple Fine Vector Flux.2 Klein 4B-----------------------------------https://civitai.com/models/2462142?modelVersionId=2768357
XCL - Transformation Workflow (FK9B)
This Workflow is intended to help create "transformation images" for the adult lewd game X-Change Life (otherwise known as XCL).No custom nodes (y'all crazy with your 200+ nodepacks)Used the following:Model: flux-2-klein-9b-fp8Clip: qwen_3_8b_fp8mixedVae: flux2-vaeLora: https://civitai.com/models/2473625
Flux.2 Klein KV (Comfyui Workflow & App)
🚀 FLUX.2 [klein] 4B + 9B AIO: Sub-Second Magic is Finally Here!Tired of waiting for your creative visions to render? Say hello to the "Speed Demon" of 2026. The FLUX.2 [klein] 4B AIO isn't just an upgrade; it’s a complete workflow revolution. We've packed the VAE, the powerful Qwen3 Text Encoder, and the UNet into one single "All-in-One" (AIO) file.Just load it, prompt it, and watch it appear before you can even take a sip of coffee.🔥 Why Everyone is Switching to Klein 4BWhile the 9B model is great for hobbyists, the 4B version is the true MVP for creators and devs. Here’s why:⚡ Sub-Second Inference: Optimized for 4-6 steps. It’s so fast it feels like real-time sketching.💼 Apache 2.0 License: Unlike the 9B (Non-Commercial), the 4B is 100% open for commercial use. Build your business on it!🧠 Qwen3 Multilingual Support: Better prompt understanding and native support for multiple languages.📉 Low VRAM King: Runs like a dream on consumer cards (8GB-12GB VRAM). Even your RTX 3060/4060 can join the party.📦 Pick Your Flavor (All-in-One Downloads)No more hunting for separate VAEs or encoders. Download the single file that fits your rig:🟢 Base-4B (The LoRA Builder) – The raw foundations for those who want to train their own styles.🟢 BF16-AIO (Maximum Quality) – Best for RTX 30xx/40xx/50xx professional work.🟡 FP8-AIO (The Daily Driver) – The sweet spot for speed and low VRAM.🔴 NVFP4-AIO (Extreme Speed) – Optimized specifically for RTX 50xx (Blackwell). Absolute insanity.🔵 FP16-AIO (Legacy Support) – For the legends still rocking GTX 10xx and RTX 20xx cards.🛠 The "Pro" Cheat SheetTo get those crisp, volumetric results seen in our examples, you must use these settings:SettingValueSteps4 - 6 (Distilled for speed)CFG Scale1.0 (CRITICAL! Higher will break the image)Sampler EulerScheduler Simple or NormalResolution1024 × 1024 (Native)🎨 Prompting Guide (Verb-Based & Volumetric)To get the most out of the Klein architecture, use descriptive, action-oriented verbs and specific lighting cues.Example 1: The Cyber-Tiger (Neon Volumetrics)Prompt: Prowling, a massive Bengal tiger advancing through a rain-slicked cyberpunk alleyway, splashing through puddles, steaming ground fog, pulsating pink and teal neon signs reflecting on wet fur, volumetric god-rays piercing through dense smog, cinematic wide-angle, hyper-realistic textures.Example 2: The Golden Lion (Atmospheric Depth)Prompt: Standing, a majestic male lion gazing into the horizon from a jagged cliffside, wind rustling through a thick mane, shimmering golden hour sunlight illuminating fur edges, volumetric atmosphere, soft bokeh mountain range, 8k resolution, National Geographic style.Ready to stop waiting and start creating?