⚡ Recommendation
You can use my custom checkpoint V2 NSFW, Lightning included:
➡️ WAN 2.2 Enhanced – Lightning Edition
for wf lightningedition long v2.1
https://civarchive.com/models/2053259?modelVersionId=2477539
Or for wf svi , no Lightning included:
➡️ SVI NSFW
https://civarchive.com/models/2053259?modelVersionId=2609141
Update: 2 or 3 KSamplers
The triple-KSampler workflow added dynamism to avoid slowdowns, but it could cause issues in certain sections, such as high brightness or other bugs. To fix this, simply enable the Double KSampler in the problematic section and change the seed.
🟣 uptade SVI LoRAs – Strengths & Weaknesses
⚡ Overview
✅ Strengths
Best solution for making long videos
Excellent transitions between video segments
Reduced degradation compared to other solutions
Strong character coherence: the model retains information from the previous video, helping maintain consistency
⚠️ Weaknesses
Weaker prompt understanding
Weaker camera understanding
Videos are less dynamic
Sometimes slow-motion effect
(can be improved with proper Lightning LoRAs and dynamic prompts)
🟣 SVI LoRAs – Download Links
⚡ Download
Note: Both LoRAs must be loaded manually in your workflow.
🟣 Suggested Lightning LoRA Combos (Optional)
⚡ Overview
You don’t have to use these Lightning LoRA combos. They are optional and allow you to fine-tune motion and degradation.
You can also use other Lightning LoRAs or assign different combos per video for more control.
🔥 Combo 1 – More Motion (Rapid Video Degradation)
High LoRA →
https://huggingface.co/Kijai/WanVideo_comfy/blob/709844db75d2e15582cf204e9a0b5e12b23a35dd/Lightx2v/lightx2v_I2V_14B_480p_cfg_step_distill_rank128_bf16.safetensors
Weight: 4
or
💜 Combo 2 – Less Image Degradation
or
or
🟢 Combo 3 – Balanced Motion / Moderate Degradation
High LoRA →
https://huggingface.co/Kijai/WanVideo_comfy/blob/709844db75d2e15582cf204e9a0b5e12b23a35dd/Lightx2v/lightx2v_I2V_14B_480p_cfg_step_distill_rank128_bf16.safetensors
Weight: 3
Low LoRA →
https://huggingface.co/Kijai/WanVideo_comfy/blob/709844db75d2e15582cf204e9a0b5e12b23a35dd/Lightx2v/lightx2v_I2V_14B_480p_cfg_step_distill_rank128_bf16.safetensors
Weight: 1.5
If you want to create several-minute-long videos while maintaining high quality, it is possible, but it will take a very long time. You just need to avoid using Lightning LoRAs and use the full model instead.
🧠 Advanced Usage Tip
You can disable global Lightning LoRAs and assign different combos per video:
Combo 1 for Video 1, Combo 3 for Video 2, Combo 2 for Video 3.
Each combo produces different motion and degradation behavior.
If you don’t use Lightning LoRAs, degradation will be even lower, but generation time will be very long.
🧠 Dynamic Prompts – Better Control & More Motion
You can use dynamic prompts to have better control and this helps to make the video more dynamic.
You just need to give this example prompt to an LLM like ChatGPT. It will be enough for it to describe the image you have and what you want as a video while keeping the same structure of the prompt in the following example.
⚠️ This will be a NSFW prompt; ChatGPT will not accept it.
You can use GROK https://grok.com, which can make NSFW prompt modifications.
For more examples of prompts with different poses, check:
Enhanced FP8 Model.
Give these prompts to GROK with the dynamic prompt structure if you want.
Example 1
(At 0 seconds: Wide shot showing a slightly overweight man casually walking down a city street, camera fixed in front, urban environment with buildings and cars.)
(At 1 second: Suddenly, a massive shark bursts from the pavement ahead, looking terrifying at first, pavement cracking, dust and debris flying, camera from side angle.)
(At 2 seconds: Medium shot from the side, the man stumbles backward in shock, while the shark dramatically slows down and strikes a comically exaggerated sexy pose, revealing large, exaggerated shark breasts, covered by a colorful bikini.)
(At 3 seconds: Close-up on the man’s face, eyes wide in disbelief, as he turns to look at the shark, small cartoon-style hearts floating above his head to emphasize his amazement, camera slightly low-angle.)
(At 4 seconds: Dynamic travelling shot showing the man frozen in the street, the shark maintaining its sexy pose, water splashes and debris still moving realistically, urban chaos around.)
(At 5 seconds: Wide cinematic shot pulling back, showing the man standing in the street, staring at the bikini-wearing shark with hearts above his head, epic perspective highlighting absurdity and humor.)
Example 2 – Anime NSFW
(At 0 seconds: The couple in a cozy bedroom, anime style, soft lighting highlighting their intimate embrace, her back arched slightly as he positions himself.)
(At 1 second: The man’s hips moving rhythmically, the head of his penis sliding effortlessly into her vagina, her body responding with a gentle, fluid motion, anime-style motion lines emphasizing the smooth penetration.)
(At 2 seconds: Her back arching deeply against him to intensify the pleasure, hips swaying with each thrust, breasts bouncing subtly, small hearts floating around them to capture the erotic energy.)
(At 3 seconds: Her face, eyes closed in bliss, a soft moan escaping, hands resting behind her head, anime-style blush on her cheeks, the air filled with a seductive aura.)
(At 4 seconds: The man penetrating her deeply, her body moving in sync with his, the bed sheets slightly rumpled, the room’s warm lighting enhancing the intimate, lustful atmosphere.)
(At 5 seconds: The couple locked in a passionate embrace, the scene exuding vibrant, seductive energy, anime style with smooth lines and soft shadows.)
findperfectresolution
https://github.com/ashtar1984/comfyui-find-perfect-resolution
rgtree
https://github.com/rgthree/rgthree-comfy
kj-node
https://github.com/kijai/ComfyUI-KJNodes
iamccs
https://github.com/IAMCCS/IAMCCS-nodes
🔧 What’s New – V2.1
The workflow is now much simpler than before.
You can create videos up to 20 seconds.
New sections can be added very easily by copying and pasting an existing section.
🎨 New Node: PainterI2VAdvanced
🔗 https://github.com/princepainter/ComfyUI-PainterI2Vadvanced
Main Parameters
color_protect
Enables or disables color drift correction.
Recommended: True.
Prevents green tint, darkening, and color distortion.
correct_strength
Controls the strength of color correction.
Higher value = stronger color correction
findperfectresolution
https://github.com/ashtar1984/comfyui-find-perfect-resolution
_______________________________________________________________________________________________
New FunVace workflow: it's extremely easy to use. If you have separate videos and want to join them with a seamless transition, use this! Thanks to Bob once again, excellent work: https://civarchive.com/models/2277993?modelVersionId=2563927
Workflow Update – Lightning Edition V2
📌 New Version Available
The Lightning Edition V2 is the latest update of my workflow.
This version is cleaner, better organized, and scene transitions are improved for smoother results.
🔧 Fixes & Improvements – Lightning Edition V2
Workflow fully reorganized and clarified for better readability.
Scene transitions improved for smoother results.
Optimized nodes and settings for easier use.
Confirmed compatibility with the Lightning Edition V2 model.
🔹 V1
Version without Super Power LoRA
A version without Super Power LoRA has been added for those who are more cautious.
If you want to try the Super Power LoRA without using the folder included in the ZIP, check the GitHub page:
➡️ GitHub – rgthree workflow
Tutorial
The workflow is divided into 6 sections:
Section 1: generates a 5-second video.
Section 5: handles transitions and interpolation.
Important notes:
Running only sections 1 + 2 will not merge the video.
To create a 15-second video, run sections 1 + 2 + 5.
Interpolation can also be enabled in section 1 if you prefer.
You can use different prompts and LoRAs for each section.
Resolution is automatically adjusted to avoid cropping.
💥 Super Power LoRA – V1
Requires the custom rgthree node:
➡️ rgthree ComfyUIInstall via git clone or the Manager, then replace the folder with the one included in the ZIP.
Allows you to save and load templates, very useful when working with multiple LoRAs, especially for WAN 2.2.
A version without Super Power LoRA is also available for those who prefer.
For the version with Super Power LoRA, see: GitHub – rgthree workflow
✅ Tips – V1
Keep the default settings for fast results.
Increase the steps if you want better quality.
Description
FAQ
Comments (59)
Hello, when I used version 2.1 to generate a 60fps video, it generated a sped-up 1-second original video. Also, Fast Groups Bypasser doesn't work correctly and needs to be manually opened. (The video quality generated by SVI is too poor; it can't even preserve faces.)
Hello! For the SVI workflow, did you use the SVI models? As for the group bypass, I don't know; there are so many reported bugs. It works fine for some and not for others. I use this workflow myself and everything works well.
it happend to me too when enable 60 fps
it just generate original video but speed up in 1s
@taek75799 Thank you for your contribution.
@bzero99 how fix，pls
@taek75799 @Naixxxx im not sure maybe OP can fix it ?
@bzero99 To be honest, I don't understand the problem. Wan outputs 16 frames. The RIFE node is used for interpolation, which means adding frames. It's set to x4 in the workflow resulting in 64 frames total. Therefore, you need to set the frame rate to 64 in the 'Video Combine' node for it to work correctly. Just make sure the RIFE node is actually performing the interpolation; otherwise, the output will remain at 16 fps.
@taek75799 i understand that but in the v2.1 workflow the RIFE you provide seem not to be working.
i have to manually add a RIFE VFI and connect from get60ffps to it then to video combine node then it work fine
@bzero99 I understand better now. So the node isn't bypassed, it just isn't working. Interesting... that's a shame because this one is faster than the one you mentioned, but at least yours actually works.
@taek75799 i like the way you use set and get node like this. it make workflow super clean, but its kinda hard to understand it fully
@taek75799 like im not understand why node set_16fps hidden beneath video4 20sec. it just work even its not enable
@bzero99 Lol, next time I won't hide anything; that way everyone can understand better and clean things up themselves, it's better that way.
For the nsfw fast move v2 q6k/q4km which workflow to use? I know svi is for svi model. But what about the lightningedition long V2.1, V2. Does long mean a model other than nsfw fast move v2 q6k/q4km?
Hello! This one: https://civitai.com/models/2079192?modelVersionId=2562360 or any other workflow
@taek75799 thank you very much.
So long is actually not the model name.
i cant seem to get the "IAMCCS_WanImageMotion" node to work. i keep hitting update but it always tells me to install it every time. what am i doing wrong?
instead of just hitting the update button, actually install it --> Custom Nodes Manager and search for "IAMCCS"
Is it normal for the video to get very dark starting from the second clip? I'm using combo 2
Hello! No, that's not normal. Most of the previews on my model page were made using Combo 2. The one here, with the missionary position preview, was done with Combo 1, which is the one that degrades the fastest. What you can do is download one of the previews from here or from my model page, save it to your desktop for example, and drag it into your ComfyUI interface. You'll be able to see if it matches your settings correctly.
@taek75799 If I want to generate a 10-second video, should I activate sections 1 and 12?
@lolbleach001584 Section 1 is active by default, so you will have 5 seconds. Section 2 allows you to add another 5 seconds; so yes, activate section 2 for a 10-second total.
I get this error when trying to run the workflow:
Cannot execute because a node is missing the class_type property.: Node ID '#961:947'
Any idea how I could fix this ?
update your comfyui
@trankhacvinh1991 workflow frontend version is 1.35.9, mine is 1.36.14. I have installed comfy 3 days ago, so it's pretty up to date
Edit: I did update to be sure but I still have the same issue
Hello! I tried something. Please try this workflow to see if the problem persists: https://www.swisstransfer.com/d/c424fc15-da6a-4b63-8f1e-29d3adb18243
@taek75799 thank you for the fast support, but I still have the same error message
@shirouste OK, what you can also do is unpack each subgraph, the ones that allow you to put the prompts
Or even simpler just try iamccs's workflow . It’s simpler and will give you the same result
https://github.com/IAMCCS/IAMCCS-nodes/tree/main/workflow_examples
@taek75799 ok I found out the issue. For some reason, IAMCSS and find-perfect-resolution did not installed. Searching for them in the extension browser yield no result.
I installed both extension manually by cloning their git repo in /custom_nodes.
Now it works :) Thanks!
@shirouste you are the champion of the people <3
im pretty new to this so i have a quick question:
In the workflow for the wan 2.2 lightning is there a way to not upscale the image for the video? you left a note saying video upscaling is time consuming, but it doesn't run for me if I turn it off. This is because the "16fps" node for videohelpersuite doesn't get the image because the the image comes form the upscaling node. So where should I get the un-upscaled image from?
ty in advance.
Hello! It’s really difficult to respond to every bug; I get them every day, especially when everything works fine on my end. But out of respect, I download the workflow (wf) and try it myself to be sure once again. I just did it: I activated the upscale node and started it. At the end, I did get a 400x400 video even though I started with 200x200 (for the test). Normally, you just need to activate the upscale node.
love your CF model and this workflow makes very good use of it. thank you very much !
two nitpicks :
- love using get/set nodes. hate seeing them hidden behind their main panels ^^ please place them next to their relevant panels :)
- those new nodes and SVI are absolutely excellent for color correction and motion - game changers compared to painterIV2 which had a good run - but I miss the fact you could setup a storyboard with FFLF ; I fail to find any way to setup a last frame in your latest workflow - or any node that would accept one ; guess that's out of the question for now ?
Hello and thank you for your feedback! Try this custom node: it has inputs for start and end images. It seems there is also a node for 3 images. It is SVI-compatible, but I haven't tried it yet to see if it works: https://github.com/wallen0322/ComfyUI-Wan22FMLF
If you can confirm that it works, I could make a simpler workflow , and I promise I'll put the "Get" and "Set" nodes right next to each other, lol!
I will give it a try today, thanks for the link !
Where's CF model ?
I'm getting mixed results with the "ultimate" node, there are multiple motion settings indeed, but which is which, I need to learn chinese, lol
the middle and last frame features on their own are actually pretty impressive ! I will try to replace IAMCCS_WanImageMotion in your new workflow with it, could be interesting
@rzxunfettered https://civitai.com/models/2053259?modelVersionId=2609141 most recent one
im using the svi workflow and i got all nodes/pieces installed (I think. It runs correctly so i assume its right), I set the no lightning cf svi for the high/low diffusion models, but when I run the workflow, the generated video has the first frame as the input image, but the rest of the video is just gray. Any ideas?
Hello! You need to use the Lightning LoRAs for the SVI model, as they are not included. Check the description for the specific combinations to use, or look at the settings within the workflow
@taek75799 I thought mine were installed since the workflow ran. I double checked and it works great now. thanks!
I’ll be releasing a new wf in a few hours. This one will be based on 3 KSamplers, which results in much less degradation and more dynamism. Thanks to bnzarev821 for reminding me of this method that I had forgotten. For those who can't wait, there is already an excellent wf using this method it will give you better results than my current SVI version: https://civitai.com/models/1866565/wan22-continuous-generation-svi2-pro-or-gguf-or-32-phase-or-upscaleinterpolate-w-subgraphs-and-bus?modelVersionId=2559451
No need for 3 KSamplers if the checkpoint has the lightning lora baked in. I hope you make an option to deactivate the 3. KSampler.
Svi 3 , group 8 has some bug , I need to bypass it to make all 12 work , also please add model offloading / free ram if I am running all groups together it's thrashing memory at very step , thanks for the amazing works !
Hello! Thanks for pointing that out, I’ve fixed it. As for the VRAM offloading, I’ll need to figure out the ideal time to trigger it when I have some free time.
I am getting pretty massive quality issues with the SVI workflow. This is all on the first generation, haven't even enabled more to extend. First, the character is not consistent with the image, not only are they not consistent but quality is degraded and the character is 'uglified' to a degree. Second, the overall video quality is just poor, there are 'random' moments of degradation and pixelation, super grainy, and movement is off.
I initially set it up using the SVI NSFW no lighting FP8 models with the recommended lighting combo for balance. Then played around with various combinations of lightning loras and models, then ultimately removed all loras and tried different models (including GGUFs) and always get the same results (to varying degrees) regardless of model and removal of all loras.
I realize this is a 'me' issue, but any thoughts on this? Anyone had something similar?
Send me the workflow that is producing poor quality specifically, the one saved after the video was generated. Thanks! https://www.swisstransfer.com/fr-fr
Ive had a ton of success whenever the initial image quality is great.
Can't install FindPerfectResolution in comfy?How can I change that node with an alternative?
Did you try installing it using git clone? Otherwise, check the description for Kijai's SVI workflow. Right after the Load Image node, I don't think it can automatically maintain the correct aspect ratio; you'll have to set it manually.
In the svi v3 workflow, the description for 'lora lightning' Combo 3 has descriptions & links that don't match. Should both high & low LoRAs be the same as the description text shows? Or should the high instead be the 1030 high in the link?
Hello! I made a mistake on the workflow posted on the first day: the link was pointing to the Lightning 1030 checkpoint. I’ve fixed it now. If the link directs you to the Civitai page for the Lightning LoRA 1030, then you’re all set.
Does that mean the description on this page is also wrong? Both in Combo 3 point to the same model.
Semi related Q: Is there much benefit between the normal 600mb rank64 1030 light high and the 15Gb f8 e4 one on this workflow to justify the bigger file?
Hello! The 15 GB file is a checkpoint; that was a mistake on my part
@taek75799 Ok thanks that makes more sense.
Also the "rank 128 bf 16" light lora returns the lora key not loaded list message for me in the console. This is with the workflow default vae and clip and the "SVI cf f8" model.
1030 and 1022 work fine with it.
@flanker151 Hello! Don't pay any attention to those messages; they are just unmapped keys due to the fact that these are Wan 2.1 LoRAs. However, they work perfectly fine on Wan 2.2.
@taek75799 Ok thanks, the workflow/model seems to work well sometimes even gives results near in quality to a full bf16 triple sampler with 5 base model steps in half the time.
The following node "object Object" is missing, so we cannot search on GitHub. Please provide a link.