Update 2026-07-05: Krea 2 Turbo v2.0 workflow added
PhotoFlow Krea 2 v2.0
Upscaling has been improved in version 2.0. There are now presets for the settings (sampler, scheduler, and denoise settings) . This allows upscaling to be performed in pixel space to achieve upscales with minimal denoising, resulting in smoother backgrounds that closely resemble the original image. In addition, the euler_ancestral_cfg_pp sampler in combination with the linear_quadratic scheduler is now recommended for high image quality. For fast generation, euler_ancestral with simple can still be used. There are also now filters to correct the colors of the final image, which can become flat due to repeated sampling.
⚠️ Note that the latest version 0.27.0 of ComfyUI (July 2027) is required to use the fast int8 models.
Update 2026-07-02: Krea 2 Turbo v1.0 workflow added
PhotoFlow Krea 2 v1.0
Krea 2 Turbo is the new king of realism and is suitable for both SFW and NSFW content. This workflow is simple, super fast, and optimized for realism.
⚠️ Note that the latest version 0.27.0 of ComfyUI (July 2027) is required to use the fast int8 models.
Update 2026-03-30: Qwen-Image-2512 Nunchaku v1.0 workflow added
PhotoFlow Qwen-Image-2512 Nunchaku v1.0
Qwen-Image-2512 offers exceptional prompt comprehension and delivers outstanding output quality. However, this comes at the cost of very long generation times due to the model's size and the large number of processing steps required. A practical solution in this workflow is to use a heavily quantized Nunchaku FP4 version, which maintains surprisingly high quality. When combined with an 8-step Lightning LoRA, generation speed improves dramatically while image quality remains excellent.
⚠️ Nunchaku requires a few additional setup steps. Please follow the instructions provided in the workflow.
Update 2026-03-21: Flux.2 klein 9B v1.1 workflow added
PhotoFlow Flux.2 klein 9B v1.1
Version 1.1 is simply a hotfix that resolves an issue with an incorrect calculation of the tile size in the Refiner step. Additionally, the EasyLoraStack has been replaced by the Power Lora Loader. The group bypass for all additional steps now also takes the previews into account accordingly.
Update 2026-03-04: Flux.2 klein 9B v1.0 workflow added
PhotoFlow Flux.2 klein 9B v1.0
Flux.2 klein 9B is a fast, 4‑step model supporting image generation up to 4 MP. The workflow starts with a recommended 2 MP base image, followed by optional face refinement and SeedVR2 upscaling to 5 MP for added fine detail. A subsequent refinement step removes SeedVR2 artefacts and enables optional upscaling up to 16 MP. A subtle touch of film grain completes the process, enhancing realism.
Update 2025-12-07: Z-Image-Turbo v2.0 workflow added
PhotoFlow Z-Image-Turbo v2.0
Version 2.0 introduces an optional Refiner and FaceDetailer, along with a recommended set of LoRAs designed to enhance photorealism and overall image quality.
⚠️⚠️⚠️
Make sure you're using ComfyUI version 0.3.76 or newer, since earlier versions contained a bug that affected the correct application of Z-Image-Turbo LoRAs.
Also ensure your SeedVR2 node is up to date, as a previous issue causing visual artifacts has been fixed.
⚠️⚠️⚠️
♥️♥️♥️
A huge thank you to all the creators of the LoRAs recommended in version 2.0. Your work makes this level of quality possible!
All images in the gallery were generated using the same set of LoRAs with identical strength settings, so in most cases no additional tweaking is required 🙂.
Special thanks to Danrisi (NiceGirls UltraReal, Lenovo UltraReal), aip0pp (PhotoReal BetterNudes), HackAfterDark (After Dark), DirtMcGirt (psxZStyle - Photography Style Lora), ConflictX (35mm Photo), and oleshevakatya (Movie Wide Angle).
♥️♥️♥️
Update 2025-11-29: Z-Image-Turbo workflow added
PhotoFlow Z-Image-Turbo
Z-Index Turbo is super fast, can generate images in 2 megapixel resolution without upscaling, and focuses on realism, so no LoRAs are required.
Update 2025-11-23: Qwen workflow added
PhotoFlow Qwen
Until now, I had been hesitant to create a workflow for Qwen because, although its image quality is outstanding, the generation times can be very long when aiming for maximum quality. The Lightning LoRA helped reduce waiting times significantly, but using it alone led to a quality drop that didn’t meet my standards. However, by combining multiple LoRAs - including the 8 Steps Lightning LoRA - with a new upscaling process using SeedVR2, I have finally achieved a level of quality that satisfies me.
I have also released a separate workflow dedicated solely to upscaling images using SeedVR2. You can find it here.
Update 2025-10-12: Chroma workflow added
PhotoFlow Chroma HD Flash
After Danrisi released the awesome Lenovo UltraReal LoRA, I gave Chroma another shot because the images looked super promising. And I gotta say, Chroma HD is amazing! Chroma needs about 50 steps in standard mode for awesome results, and even with an RTX 5090, that's too long of a wait for me. Fortunately, there is a Chroma Flash LoRA that achieves great results with just a few steps. With CFG 1.0, the time per step is halved again!
Update 2025-08-23: Wan 2.2/2.1 T2I workflow added
PhotoFlow Wan T2I
Wan 2.1 and Wan 2.2 are video models that can also be used to create high-resolution images. Due to the impressive quality, I decided to create a dedicated workflow for them. Wan 2.2 was released during the development of this workflow. Fortunately, Wan 2.2 is even better and can be used like the Wan 2.1 model, including compatibility with most Wan 2.1 LoRAs. Style LoRAs, in particular, work very well. SageAttention with Triton is supported but not required, and the Lightning LoRA is used to ensure fast generation of high-resolution images. Instructions on how to set up and use the workflow are included within the workflow itself.
PhotoFlow XL
This SDXL ComfyUI workflow has been developed to generate realistic photos with a focus on people.
The version 2.0 is not restricted to the use of the DMD2 Speed LoRA. The workflow works just as well without it ♥️.
The default settings and comments in the workflow are geared towards DMD. Accordingly, you only need to adjust the CFG, sampler, scheduler and steps for a non-DMD use 🙂. These values are a good start for a generation without the DMD LoRA:
KSampler
CFG: 3.5 (2.5 - 5.5)
Sampler: dpmpp_sde_gpu
Scheduler: normal
Steps: 25
FaceDetailer
CFG: 3.5 (2.5 - 5.5)
Sampler: dpmpp_sde_gpu
Scheduler: karras
Steps: 18
Recommended Checkpoints
The following checkpoints are particularly suitable:
SFW
epiCRealism XL: https://civarchive.com/models/277058?modelVersionId=1920523
HelloWorld XL: https://civarchive.com/models/43977/leosams-helloworld-xl
Juggernaut XL: https://civarchive.com/models/133005/juggernaut-xl
RealVisXL 5.0: https://civarchive.com/models/139562?modelVersionId=789646
NSFW
Jib Mix Realistic XL: https://civarchive.com/models/194768?modelVersionId=1809082
DonutsDeliveryMix: https://civarchive.com/models/605283?modelVersionId=1739105
Lustify: https://civarchive.com/models/573152?modelVersionId=1094291
Analog Madness SDXL: https://civarchive.com/models/408483/analog-madness-sdxl-realistic-model
Big Love XL2: https://civarchive.com/models/897413?modelVersionId=1464711
XXX-Ray v1.1: https://civarchive.com/models/1064836?modelVersionId=1305584
NSFW (DMD included)
GonzaLomo (Unity/Real XL DMD): https://civarchive.com/models/1513492/gonzalomo-xlfluxpony
Lustify ENDGAME DMD2: https://civarchive.com/models/573152?modelVersionId=1099200
XXX-Ray DMD2: https://civarchive.com/models/1064836?modelVersionId=1624818
Installation
ComfyUI Manager Installation
https://github.com/ltdrdata/ComfyUI-Manager?tab=readme-ov-file#comfyui-manager
Installation of Missing Nodes
Missing nodes in this workflow are highlighted in red. To install them, open the ComfyUI Manager and click on the “Install Missing Custom Nodes” button.
⚠️‼️ComfyUI-KJNodes needed‼️⚠️
It it may happen that the ComfyUI-KJNodes are not included in the list of missing nodes. Simply install the ComfyUI-KJNodes via ComfyUI Manager > Custom Nodes Manager.
After the installation, restart ComfyUI and refresh the browser page.
Further information and tips on configuration can be found in the workflow in the INFO node.
TL;DR
The workflow was developed and tested with a NVIDIA RTX 4080 (16 GB VRAM) and 96 GB RAM.
Leave a positive review, if you like the workflow 😊.
Description
Version 2.0 introduces an optional Refiner and FaceDetailer, along with a recommended set of LoRAs designed to enhance photorealism and overall image quality.
FAQ
Comments (33)
Not work with z-image "SAMLoader stat: path should be string, bytes, os.PathLike or integer, not NoneType"
Right click > Fix node (recreate)
@qek I tried your suggestion but it did not work. I unpin and then expand the node, I can't select any models under the model/sam folder. BTW, thanks for the workflow, works pretty good even when detailer is disabled to get around the problem.
@Cowgary Maybe the model wasn't installed properly. You can search for "sam_vit_b_01ec64" in the "Model Manager" of the "ComfyUI Manager" and install it.
@ai839 Thanks for the quick respond. I'm a freaking idiot because after using the Model Manager to download, I noticed my manual install folder is "sam" and not "sams". Problem is now resolved.
FaceDetailer is a great addition to the V2, it seems to enhance likeness when using a lora (tough I did tweak the settings). The refiner seems to consistently make the image worse than the base generation tough.
I agree, the Refiner step in the current configuration makes the image worse in most cases. It performed better than an SDUltimateUpsale in my tests during the development of this workflow. I will find a better solution in future workflows.
@ai839 Ah no worries, its still a great wf, my favorite so far. I just added Detail Daemon for the base generation and it's looking good! Oh and also Conditioning Noise injection for some more variety.
I tried with another workflow and I have the same issue. ZiT gives very clean results and the detailer noticeably degrades the sharpeness and cause distortions (especially with the eyes and the bokeh) even with a low strength. That said, ZiT is very powerful for prompts, I think it's best to work more on this rather than using a detailer. This also applies to many LoRa I've tested.
Qwen workflow doesn''t work for me at all. Model switch node is circled in red and WAS Node Suite shows Import Failed message.
The original author of the WAS Node Suite no longer maintains the project, but you can resolve this by installing "WAS Node Suite (Revised)".
I gave this a shot and, like your other workflows, it is quite impressive. The preview images aren't loading, though that might mean I need to activate something that isn't already activated by default.
I was thinking of trying to adjust this for image to image, but I am in the dark about a lot of the nodes in this flow. Any thoughts?
You mean the preview in the sampler nodes? Just open the ComfyUI Manager and set the "Preview method" to "Auto". In the newest ComfyUI version the setting moved to the ComfyUI settings. Search for "Live preview method" and set the value to "auto".
Oh, you’re talking about image-to-image refinement? Then just add a Load Image node and a VAE Encode node, and use its latent output for the sampler instead of the empty latent in the default workflow. Make sure to lower the denoise value. The lower it is, the fewer changes you’ll see in the result.
@ai839 Oddly enough, the preview node came to life on its own. Very strange that it wasn't working and then suddenly did. But, that's comfy for ya.
I will go ahead and add the node - just gotta remember how. I haven't added to a workflow in ages, and my memory sucks!
Okay. I figured out why the preview images weren't showing up. It's the LoRa loader. I had some of the LoRa's loaded with a "ZIT_*Name*" and they weren't loading. If the LoRas don't load, the pre-images do not show.
it works nicely. could you implement adding a node for image reference?
Using an image reference as denoised latent?
Any thought of revisiting the older workflows and updating them? I still really love the WAN one--the quality is excellent and I love the images I get from them (though it does suffer from some "same face" issues). It's still using an older Smartphone Snapshot lora even though there are newer ones, for example. Or perhaps looking at updating it to use detailers or seedvr like you have done for Z image.
Thanks for your efforts!
I don't know if I'll find time for it. I agree with you that the older workflows could need an update. Time will tell...
very nice workflow. only thing i had to change was a little upscaler settings to fit my vram. other than that, best workflow on civit by far
Late answer: I am curious about your hardware and settings. I would like to integrate tips into workflows about which settings work for SeedVR2 for how much VRAM and RAM. My defaults are for my setup with an RTX 5090 with 32 GB VRAM and 96 GB RAM. These are not actually good defaults, as very few people have such a setup.
@ai839 basically once i made sure IO block swapping was enabled and set blocks to swap to 14, it works fine on my 12gb 3060. few small tweaks to the upscaler settings after testing made it run perfectly fine without loosing any quality by changing encode tile size.
@RPxGoon So you activated tiling? For encoding and decoding? 1024 tile size?
Great workflow! Could we have an Image to Image Version?
No plans for that until now. Can you explain what's your understanding of an image to image workflow?
I simply can't install res4lyf, no matter what, I lowered the security settings for the Comfy Manager, tried to install manually, nothing, can someone please help me?
Unfortunately, I overlooked your comment. If the problem still persists, you can also install RES4LYF manually: https://github.com/ClownsharkBatwing/RES4LYF?tab=readme-ov-file#installation
Are you talking about the Manager integrated in ComfyUI or the older ComfyUI-Manager (https://github.com/Comfy-Org/ComfyUI-Manager)? For me only the older works.
That WF looks great, I am using ZiT but I am getting same results as for ZiT - cmfy template wf when I disable upscaling and seedvr2 ( just to get faster outputs). Seems like face detailer is doing nothing. First two nodes down for images I get empty after finished generation. But I might do also something wrong so.. hard to say. I got only 64GB DDR4 and RTX4070
This are BY FAR the best workflows on Civit! I don't use any others anymore... THANK YOU!
I'm having a problem with the XL workflow. The workflow stops—without any errors—after Base Gen. FaceDetailer is outlined in red, but I don't understand why. The settings are all logical, so it can't be my hardware. Do you have any ideas?
Edit: Ah, it's the Scheduler Node. I deleted the node and used the selection in FaceDetailer, and then it worked. No idea why.
@CupRunethOverWithLSD Yes, the workflow is quite old and not compatible anymore. Your solution is correct.
@ai839 Can you perhaps go through some of the older workflows and update them? Overall this seems so good but AI gen moves so fast and so the older ones get outdated or stop working.
@ptangwai700 I would love to not only get everything up and running, but also improve it, as I have learned a lot over the months and years and, as you said, development is very fast and there are lots of new things coming out that can also be combined with older models. However, it is completely unrealistic that I will find the time for this. I can't promise anything, but maybe I will at least manage to fix the technical errors in the near future.