Master Flux ControlNet Canny in ComfyUI with Lightning-Fast Setup!

Unlock the Flux controlnet Canny in ComfyU. Whether you're using a high-end system or just getting started with basic hardware, I'll walk you through optimizing your workflow, from downloading the right models to producing stunning images with precision through Flux controlnet Canny in ComfyUi . Don't miss out on how you can turn your ideas into visuals faster and better than ever!
In this video:
I’ll show you how to set up and run the optimized Flux and Canny controlnet for AI-generated artwork. We’ll cover:
Downloading and installing necessary models
Step-by-step guide to setting up the workflow for Flux controlnet Canny
How to fine-tune settings for maximum accuracy and creativity in your output
A comparison of your input image with the final result, transformed from the canny model map.
Buy me a coffee ☕ :
/ shop
I promise to make it worth your while with even better tutorials
Links:
Download the Flux.dev FP8 : huggingface.co...
Download the Flux controlnet canny V3 : huggingface.co...
Download new text encoder for flux: huggingface.co...
Download this workflow: drive.google.c...
Access Replicate for controlnet canny model on the cloud: replicate.com/...
Watch my previous video on ComfyUi full installation: • Install FLUX locally i...
Watch my previous video on Flux GGUF models installation: • Install Flux locally i...
Watch my previous video on How to make Flux images of yourself: • 💥Create Epic AI Photos...
.
Complete Guide for Beginners (watch videos below on by one) :
1-Install ComfyUi and Flux Locally : • Install FLUX locally i...
2-Guide for Low-end systems for Flux : • Install Flux 1.0 Dev 2...
3-How to Create ai images with your own face : • I Tried Flux Lora trai...
Thank you for watching! If you found this video helpful, don’t forget to give it a thumbs up, subscribe to the channel, and hit the notification bell so you don’t miss any future content. See you in the next video!👇🏻
/ @jockerai

[@Jockerai]

click the link below to stay update with latest tutorial about Ai 👇🏻 : www.youtube.com/@Jockerai?sub_confirmation=1

[@AlphaGirth]

Fantastic tutorial and workflow. Much appreciated.

[@KaganParlatan]

Even though I know how to install all, I like the way how you explain, very simple. Good work, earned my subs.

[@Jockerai]

@@KaganParlatan you're very welcome my friend ✨ thanks for sharing your opinion it was very uplifting ✌🏻

[@cocosloan3748]

I tested it. It works. TY for workflow 👍

[@relexelumna5360]

Thanks for such a valuable detail information speciallly "the Links" 🙌

[@Jockerai]

You're very welcome my friend

[@hoomanmcak4957]

Thx for your amazing content …. Persian accent is the best every time 🔥🇮🇷

[@Jockerai]

@@hoomanmcak4957 thank you so much bro ❤️🔥

[@giuseppedaizzole7025]

Hi there...i wonder if could be possible to add a lora in this confyui workflow. thanks

[@CENTRALFCP]

great video and links to download all

[@Jockerai]

@@CENTRALFCP thank you ✨😉

[@Cainra11]

Hi, thank you for the content and the free tutorials! However, I encountered a problem-my original photo couldn't properly generate the Canny edge detection, so the final image turned out to be something unrelated. Is there a way to resolve this issue? I’d appreciate your help!

[@julayen]

Thanks joker, your video help me a lot. With 3d software i have already generate a depht map. I think its easy to use it with depht instead canny. But can i tune this workflow and bypass the detph generation and put directly mine instead. Or comfy need to do a new one ?

[@sammehrani-q6h]

dadash damet garm

[@Jockerai]

@@sammehrani-q6h you are welcome bro🌱❤️

[@Aibyda]

I m getting error mat1 and mat2

[@motionbox4]

i found some error everytime, this time controlnet loader not showing any options in loader file..only updated flux 8 dev...why is this... every time i not able to achieve...after spend some time...

[@motionbox4]

showing undefined, not able to see any options under controlnet..

[@ralf.starke]

Nice, but why do I get only comic stlye pictures? Canny works with Juggernaut great, something seems wrong with Flux.

[@Jockerai]

@@ralf.starke change the prompt

[@ralf.starke]

@@Jockerai Thanks. My prompt starts with "a photorealistic..." so it should be fine? In your video the woman is also only some painting and not photorealistic. Are you able to get some photorealistic results like the input load image?

[@Jockerai]

@@ralf.starke yes of course. Just play with parameters like flux guidance. Increase it to make ai more adherence to your prompt . Also play with Apply Flux Controlnet strength.

[@ralf.starke]

@@Jockerai Thank you, I will try. Or maybe someting with my installation is wrong.

[@RenjithRS]

Xlabs sampler is giving out of memory errors. Any alternatives available?

[@Jockerai]

@@RenjithRS what is your system configuration?

[@Aibyda]

You should have 24gb graphics card to use controlnet

[@maikelkat1726]

He, I want to test the workflow, but its not available? could you add it to the description? thanks.,

[@Jockerai]

Description updated now

[@hansegon496]

Great, but where is the link to the text encoder? Can yo please share it?

[@Jockerai]

I added it in description

[@CsokaErno]

Could you share the link of the DualClip loader you mention in the video? Thank you. Sorry for a tiny note but if you keep more distance between the floating windows, it would be easier to follow the spagettis. Thank you, anyway :)

[@Jockerai]

I added it in description

[@CsokaErno]

@@Jockerai Yes, but there are four files which starts with ViT-L-14-TEXT... And it is not a picture with text :)

[@Jockerai]

@@CsokaErno I showed you in previous video which one I'm using HF.safetensors

[@CsokaErno]

@@Jockerai I don't know which video you mean. Okay, never mind. As far as I see other people have the same problem here :) Thank you, anyway.

[@Jockerai]

@@CsokaErno kzbin.info/www/bejne/i2W1fZ16gNeAaacsi=ytcEyo7RZ_L3gN6o

[@pavilionbintaro]

Hi @ jockerai i use the same detail why my result look pixelated in comfy ui

[@Jockerai]

@@pavilionbintaro make sure all the nodes in your workflow is as same as mine. And one of the most important ones is the Flux guidance node which if missing you'll get pixelated images

[@KINGLIFERISM]

Just so you know it wont look good because controlnet on flux isn't implemented well. You will see. Do not take my word for it.

[@Jockerai]

@@KINGLIFERISM it's not true I make tons of images with controlnet canny using Flux just like in this video, even my channel banner and for my customers. additionally you can see on replicate there are online website which allows you to use canny depth etc with flux. you can search for it

[@costatattooz840]

is 12 GB VRAM (RTX 3060) enough? i have also 32 GB RAM. using comfyui on archlinux

[@Jockerai]

yes it is great

[@costatattooz840]

@@Jockerai You are the best man! Got a new sub! Howerver can i ask a few things:? 1) I launched nvidia-smi and i saw that i have another 1,5 gb of vram to spend.. There is something better i can download? like flux model better in quantization or clip models? 2) what does really do flux guidance value and strenght value?

[@burskyozbekov]

there is no workflow...:(

[@yiluwididreaming6732]

Just follow the video bloke.......easy peasy.... 😉

[@Jockerai]

you can follow the video and do them one by one, it's better for you to better understanding of ComfyUi and learn much more. Besides you support me by more watching my videos which is very appreciable😉✨

[@Dangerkabaap]

Workflow?

[@Jockerai]

@@Dangerkabaap I'll put it in the description

[@HiProfileAI]

@@Jockerai not in the description?

[@Jockerai]

@@HiProfileAI Description updated

[@HiProfileAI]

@@Jockerai Thanks

[@luridape7486]

I've been unable to get this running without an OOM. 48GB DRAM, i5 9400k, RTX 3060TI 8GB. Tried flux1-dev-Q4_0.gguf, t5-v1_1-xxl-encoder-Q4_K_S, 400X400 image. Is there any way to lower the requirements or am I out of luck? [edit: Running Ubuntu 24.04 headless to leave all VRAM available]

[@sammehrani-q6h]

dadash damet garm