March 21st, 2026
Added Auto strength node
Instead of applying one flat strength across the whole network and guessing if it's too much or too little, it reads what's actually in the file and adjusts each layer individually. The result is output that sits closer to what the LoRA was trained on, better feature retention without the blown-out or washed-out look you get from just cranking or dialing back global strength.
One knob. Set your overall strength, everything else is handled.
The manual sliders are optional choice for if you don't want to use the auto strength node! but I 100% recommend using the auto-strength node
March 10th, 2026 (Fixed refresh bug)
A bug was causing the new added Lora not to appear upon page refresh has been fixed.
Also for more of my work and other node packs you can see them at my GitHub Page
Here: https://github.com/capitan01R
and if you would like to support my work you can buy me a coffee at this link :)
March 7th,2026 :
There was a bug with the Lora stack node when refreshing the browser, but now it's fixed!
Been using Z-Image Turbo and my LoRAs were working but something always felt off. Dug into it and turns out the issue is architectural, Z-Image Turbo uses fused QKV attention instead of separate to_q/to_k/to_v like most other models. So when you load a LoRA trained with the standard diffusers format, the default loader just can't find matching keys and quietly skips them. Same deal with the output projection (to_out.0 vs just out).
Basically your attention weights get thrown away and you're left with partial patches, which explains why things feel off but not completely broken.
So I made a node that handles the conversion automatically. It detects if the LoRA has separate Q/K/V, fuses them into the format Z-Image actually expects, and builds the correct key map using ComfyUI's own z_image_to_diffusers utility. Drop-in replacement, just swap the node.
Repo: https://github.com/capitan01R/Comfyui-ZiT-Lora-loader
If your LoRA results on Z-Image Turbo have felt a bit off this is probably why.
Description
v1.2.0
added Multiple Lora Load and remove
FAQ
Comments (20)
Nice! Small bug report: The Lora stack loader loose all loras upon re-loading the workflow from an image created with it.
EDIT: Typo
Yeah I second this, it's forgetting the loras, but otherwise seems like a great node!
yes true I just fixed that, it was the js file :)
@FairyDave fixed!
@Capitan01R Amazing, thank you I'm gonna give it a go now :D
@FairyDave it's still buggy, but I'm currently working on fixing it fully.. I apologize edit : now it's fully fixed!!!
@Capitan01R No need to apologise man, I'm grateful for the effort you put in. It's looking good in my wf, so thank you :D
@Capitan01R Thank you! (I agree with FairyDave above). Great stuff!
I think you forgot to serialize the settings in your code? they are lost after a browser refresh (you can put them in the property panels so they are serialized automatically on change)
fixed
indeed, thanks!
another thing: browser refresh doesn't refresh your lora list, it requires a full comfyui restart..
@yodaSteelBeans I will work on a fix I did not notice that during the runs.. apologies.
it's alright, took me a while to fix that for my own loaders.. it should be in the PY file if memory serves, I can send you my source code if you can't find the fix. Great fix btw
@yodaSteelBeans fixed now!
works great, thanks
Hey, I think there may be a bug in v1.2 where it doesn't pick up new loras added to the folder, even after hittin r to refresh. My be a skill issue my side, but I think it's a bug :D
Yes true that bug has been reported, I was just quite busy the past few days but I will work on it tonight and hopefully will push the fix after that in this same update this way it’s not a new version.
now fixed!
@Capitan01R quick work! thank you so much :)