Thank you for this plugin. It is extremely helpful. As for additional features, here's my request: 1. Ability to add custom controls to the PS Plugin UI. I mostly use modified version of your workflow. I have added a custom branch for IC-Light and it's extremely helpful. But it has several settings that I currently adjust directly in comfyui. It would be great to have a way to adjust more than one numeric setting directly in the UI. Would be even greater to have other types of controls (e.g. dropdowns with custom elements) 2. I often faced an issue with masks not being passed to the workflow. For example, even after a fresh start of both PS & comfyui, I create a custom mask and start the process, but in the workflow I still see an old mask. However, I don't have reliable steps to reproduce it. It just occurs randomly, like you do it several times and it works fine and then you just run in one more time and it happens. So I couldn't verify if you already fixed it in the latest version. 3. Randomize button requires an extra click to reach to, prompt textarea is too small, it would be good if seed == -1 worked as a random seed, it might be benefitial to have a "batch" option. For example, you batch 10 generations and then it runs 10 times with different seeds and puts 10 generations on separate layers.
Hello Nima!!! Cool plugin. I just tried it. I have one request. I took an image of a girl with her eyes closed. And in the mode in Paint I set the hint OPEN EYES. Everything turned out great, but on the generated layer, the area where the eyes are drawn, you can see that the resolution of the image in this place is much lower. On your git channel, there is a way to avoid this problem by pre-loading a workflow for redrawing portraits. But this does not take into account all the remaining cases. For example, in the in Paint mode, there will be a need to depict, for example, a building. This created building will have a low resolution. Is it possible to change the in Paint mode so that you can create any required element with the resolution of the entire edited image? There are a lot of examples of this on the Internet. I even tried replacing that block in Paint using a custom Inpaint Crop node, but it didn't work
There is currently a bug where the generated mask has raised "pure black" from 0,0,0 to 1,1,1. I believe Nima is working on this as there is an open item on his Github. If fixed, then these nodes should work well with Inpainting Crop nodes (at least the ones I've used). I now use Inpaint Crop and Stitch nodes (there are others) because it's the closest I have found to inpainting in A1111. Using your example, you could select a region that is 300 X 400 PX around a building in a much larger image that is 2600 X 4200. In that node, you set a % increase or # of pixel increase (beyond the mask) to the image "for context" before generation. The nodes crop to that region (and enlarges it if you want for added detail), use the mask for generation, and then stitch it back together (downscaling it if you had enlarged the image first). NIma's UI already allows us to add the "enhanced inpainting" workflows. Just waiting on this one fix and I think we'll have it. Hoping!
Feature request about the target size : it would be useful to have presets and aspect ratios similar to the latest versions of A1111. The size should be divisible by 8 for the VAE to perform well.
Thank you for your great work. The previous version was simpler for me. also the new workflow json file was broken in some parts (KSampler in img2img Line CN) and i had to fix it. maybe it was something else that i couldn't understand.
Great job!!!!!! next features: face swap, dress change, Ai skin retouching (using frequency separation) one click background remover (better than Photoshop's own bG remover) something like remove,bg
I loved how simply and easy was your first version. Just one node to send PS to Comfy. It is possible to use it this one like that? I prefer to conect PS with my own workflows 🤔🤔
@@nekodificador but you have to know that it's not possible to control it inside of Photoshop! Instead you can use new version but make your own workflow! No need to use default workflow at all!
Hi, thanks for providing this amazing plug-in.😊 Just tested it on different versions of Photoshops. A little suggestion, you should change the "minVersion" to 25.0.0 in the manifest file of the plug-in since there are several new PS API structures that couldn't be run at the old PS versions.
I am experiencing connection issues between Photoshop (PS) and COMFYUI. Additionally, is there a way to prevent images from being sent at a low default resolution?
Firstly, thank you to the author for providing the plugin. I'm currently facing connection issues between Photoshop (PS) and COMFYUI. I've tried troubleshooting by disabling a translation plugin as suggested in other posts, but the problem persists. The most significant issue I'm encountering is that images sent from PS to COMFYUI are being compressed to a very low resolution, and adjusting the settings does not seem to change this. Could the author consider updating the plugin to allow for the default send resolution to match the current canvas or rectangular selection resolution in PS? This would make the plugin more usable while we work on resolving the connection issues.
Or, could the author provide some temporary solutions to address this resolution issue? I tried looking into the code to adjust the file transmission values, but was unsuccessful.
Great tool thanks, As for updates I would like to see a quick template update, so I can Add more custom drop downs for models and steps in that section, as it would just be a quick setup in comfyUI and then all I have to do is just have them ready in the drop down boxes in photoshop plugin :) Also the feathering seems to be bugged and would be good if there was an option you could tick that asked you if you want to feather the selection edge when you clicking send to layers and it prompts you with the amount then sends it to the photoshop layers. Thanks for this epic plugin!
About the second one , you already have an option to disable it in setting, simply disable "fix mask edge" checkbox on image proccessing section in settings About the first part i really didn't understand what you mean, we can already create our own custom workflow and add it into drop down
@@NimaNzrii I'm not to sure how to go about the first part and would be quite useful to get a video on it just to breakdown like how your workflow works, if possible as I'm sure a lot of others would find it useful. As for the second it would be good to get a pop to define the level of feather that's all not massive issue though, As I have been manually doing it, Partly because sometimes the fix mask edge I found can have a bit of image bleed.
When loading the graph, the following node types were not found: workflow/math 4 Slider ( a slider value ) Nodes that have failed to load will show as red on the graph. it says like that
@@NimaNzrii آره پلاگین و نودها و ورکفلو رو آپدیت کردم ولی مثل همون نسخه قبلی خروجی سیاه میده و توی این آپدیت جدید داخل تصویر سیاه مینویسه نو دیتا ، بعد یه بار دیگه رندر بزنم همونو ایمیج2ایمیج میکنه متن نو دیتا عوض میشه مشخصه که تصویر داره میره و میاد ولی نمیدونم چرا از لایه های فتوشاپ نمیخونه هیچی. ورژن فتوشاپم هم 24
I modified the nodes in comfyui, and I want to change the name of each group, or add some groups, but there is no change in photoshop, how can I do this?
May I ask you if using this plugin bypass the photoshop autogenerative rules? EG, they won't generate a Honda or NSFW stuff. Not saying that I need to generate those. But we must not be restricted as Photoshop is just a tool. Not a police to tell me what not to do. And you know what? Why did I ask you that? I can find out myself. lol . Thanks for the video.
I changed the workflow you gave me, and renamed each group in Chinese. When I loaded the workflow again, I couldn't find these groups in the UI, but I could click to generate them. I hope the next version will be improved. Thank you for bringing us such a good plugin.
Nima how can I run this on my laptop as a server with --listen args I connect to my local newtwork from my Apple Mac PC. I use photoshop my mac, and stable diffusion on my windows. It works for web browser comfyui but photoshop it will not work! 4:27 ?
There is also problem with default instructions with the node titled workflow/math 4 slider (a slider value), node is missing from comfyui manager! work around with right click the existing node and click convert node to widget>convert strength to widget until there is a fix
I got the same error. When loading the graph, the following node types were not found: workflow/math 4 Slider ( a slider value ) Nodes that have failed to load will show as red on the graph.
Please create a tutorial on editing photos in photoshop using comfy - retouching, color grading, relighting Please Btw awesome video Always eager for next videos ❤❤❤❤
I usually use Chinese comfyui and Chinese PS, but there will be problems if I don't modify it into English. Chinese comfyui can't recognize parameters such as width, height and seed, and Chinese ps can't recognize selection in comfyui with the local redrawing function. How can I solve this problem
@@NimaNzrii Download this plugin (AIGODLIKE-ComfyUI-Translation) and click switch locale in the bottom right corner, and comfyui becomes Chinese, and then ANYTHING EVERYWHERE in your node's original image about width, height, seeds, etc. That's the first question! The second problem is that the ps language is changed to Chinese, and the version of PS is whatever. It doesn't matter if I use the version of 2024 and 25.10. In fact, when PS is Chinese, no matter whether comfyui is Chinese or English, MASK nodes like #1250 #597 are blank, and the selection of PS cannot be obtained. Therefore, both in-Paint and out paint are invalid IN PS, and everything is normal when ps is in English mode and comfyui is also in English mode. However, such a way makes our foreign users too uncomfortable, I hope you can help solve it, thank you
Hi, I have some problems and it would be really helpful if you have any clues on how to solve them. At 6:44 in the Node Photoshop ComfyUI plugin, mine shows "no data." At 1:24 in Presets, both of the lists show "couldn't load." Thank you very much!
Great work - have already started adding my own sdxl "processes" to the base flow - so much potential. No need to reconnect noodles just send your image to a different process from the plugin
When you run the node it will automatically install it for you! No matter which ps version you have Check your ps to see if it worked or not But if it didn't work just use zxp uxp installer like the video
Thank you for this plugin. It is extremely helpful.
As for additional features, here's my request:
1. Ability to add custom controls to the PS Plugin UI. I mostly use modified version of your workflow. I have added a custom branch for IC-Light and it's extremely helpful. But it has several settings that I currently adjust directly in comfyui. It would be great to have a way to adjust more than one numeric setting directly in the UI. Would be even greater to have other types of controls (e.g. dropdowns with custom elements)
2. I often faced an issue with masks not being passed to the workflow. For example, even after a fresh start of both PS & comfyui, I create a custom mask and start the process, but in the workflow I still see an old mask. However, I don't have reliable steps to reproduce it. It just occurs randomly, like you do it several times and it works fine and then you just run in one more time and it happens. So I couldn't verify if you already fixed it in the latest version.
3. Randomize button requires an extra click to reach to, prompt textarea is too small, it would be good if seed == -1 worked as a random seed, it might be benefitial to have a "batch" option. For example, you batch 10 generations and then it runs 10 times with different seeds and puts 10 generations on separate layers.
great suggestions! will definitely add them on next version!
Hello Nima!!!
Cool plugin. I just tried it.
I have one request. I took an image of a girl with her eyes closed. And in the mode in Paint I set the hint OPEN EYES. Everything turned out great, but on the generated layer, the area where the eyes are drawn, you can see that the resolution of the image in this place is much lower. On your git channel, there is a way to avoid this problem by pre-loading a workflow for redrawing portraits. But this does not take into account all the remaining cases. For example, in the in Paint mode, there will be a need to depict, for example, a building. This created building will have a low resolution. Is it possible to change the in Paint mode so that you can create any required element with the resolution of the entire edited image? There are a lot of examples of this on the Internet. I even tried replacing that block in Paint using a custom Inpaint Crop node, but it didn't work
There is currently a bug where the generated mask has raised "pure black" from 0,0,0 to 1,1,1. I believe Nima is working on this as there is an open item on his Github. If fixed, then these nodes should work well with Inpainting Crop nodes (at least the ones I've used). I now use Inpaint Crop and Stitch nodes (there are others) because it's the closest I have found to inpainting in A1111. Using your example, you could select a region that is 300 X 400 PX around a building in a much larger image that is 2600 X 4200. In that node, you set a % increase or # of pixel increase (beyond the mask) to the image "for context" before generation. The nodes crop to that region (and enlarges it if you want for added detail), use the mask for generation, and then stitch it back together (downscaling it if you had enlarged the image first).
NIma's UI already allows us to add the "enhanced inpainting" workflows. Just waiting on this one fix and I think we'll have it. Hoping!
Photoshop과 ComplyUI를 연결할 수 없습니다. Photoshop에는 생성 창이 있지만 작동하지 않고 연결되지 않습니다. 무인입니까?
Superb job man! Thank You!
Amazing, thanks for making ComfyUI so much better
SDXL please
Feature request about the target size : it would be useful to have presets and aspect ratios similar to the latest versions of A1111. The size should be divisible by 8 for the VAE to perform well.
photoshop Not Connected
Thank you for your great work. The previous version was simpler for me. also the new workflow json file was broken in some parts (KSampler in img2img Line CN) and i had to fix it. maybe it was something else that i couldn't understand.
Thank you @NegariaDesign what made the prv version simpler for you?
Great job!!!!!!
next features:
face swap,
dress change,
Ai skin retouching (using frequency separation)
one click background remover (better than Photoshop's own bG remover) something like remove,bg
hey imran,
Noted!
It seems that it is impossible to achieve two diagram guidance, one for IPA and one for controlnet
photoshop Not Connected
I loved how simply and easy was your first version. Just one node to send PS to Comfy. It is possible to use it this one like that? I prefer to conect PS with my own workflows 🤔🤔
@@nekodificador the old method still working but the node name change to "🔹 Photoshop remote connection"
@@nekodificador but you have to know that it's not possible to control it inside of Photoshop!
Instead you can use new version but make your own workflow!
No need to use default workflow at all!
@@NimaNzrii Ok I tried and 100000% this is the best thing for retouchers in YEARS. Works like a charm
Can you connect the plugin to a non local PC running comfy?
Can a Chinese interface be added?
Hi, thanks for providing this amazing plug-in.😊 Just tested it on different versions of Photoshops.
A little suggestion, you should change the "minVersion" to 25.0.0 in the manifest file of the plug-in since there are several new PS API structures that couldn't be run at the old PS versions.
I am experiencing connection issues between Photoshop (PS) and COMFYUI. Additionally, is there a way to prevent images from being sent at a low default resolution?
Firstly, thank you to the author for providing the plugin. I'm currently facing connection issues between Photoshop (PS) and COMFYUI. I've tried troubleshooting by disabling a translation plugin as suggested in other posts, but the problem persists. The most significant issue I'm encountering is that images sent from PS to COMFYUI are being compressed to a very low resolution, and adjusting the settings does not seem to change this.
Could the author consider updating the plugin to allow for the default send resolution to match the current canvas or rectangular selection resolution in PS? This would make the plugin more usable while we work on resolving the connection issues.
Or, could the author provide some temporary solutions to address this resolution issue? I tried looking into the code to adjust the file transmission values, but was unsuccessful.
Great tool thanks,
As for updates I would like to see a quick template update, so I can Add more custom drop downs for models and steps in that section, as it would just be a quick setup in comfyUI and then all I have to do is just have them ready in the drop down boxes in photoshop plugin :)
Also the feathering seems to be bugged and would be good if there was an option you could tick that asked you if you want to feather the selection edge when you clicking send to layers and it prompts you with the amount then sends it to the photoshop layers.
Thanks for this epic plugin!
About the second one , you already have an option to disable it in setting, simply disable "fix mask edge" checkbox on image proccessing section in settings
About the first part i really didn't understand what you mean, we can already create our own custom workflow and add it into drop down
@@NimaNzrii I'm not to sure how to go about the first part and would be quite useful to get a video on it just to breakdown like how your workflow works, if possible as I'm sure a lot of others would find it useful.
As for the second it would be good to get a pop to define the level of feather that's all not massive issue though, As I have been manually doing it, Partly because sometimes the fix mask edge I found can have a bit of image bleed.
Thank you for plugin. As for A1111 it would be super nice to be able work with masks and with selection
When loading the graph, the following node types were not found:
workflow/math 4 Slider ( a slider value )
Nodes that have failed to load will show as red on the graph.
it says like that
درود بر تو 💚💚.
ولی برای من تصویر خروجی سیاه هست و توی این آپدیت جدید هم noData میاد. لطفا issues شماره 72 گیت هاب رو چک کن. ویدیوش رو گذاشتم
سلام ، نود و پلاگین رو اپدیت کردید؟ چون ایشو شماره ۷۲ نسخه خیلی قدیمی ای از نود و پلاگین هست
@@NimaNzrii آره پلاگین و نودها و ورکفلو رو آپدیت کردم ولی مثل همون نسخه قبلی خروجی سیاه میده و توی این آپدیت جدید داخل تصویر سیاه مینویسه نو دیتا ، بعد یه بار دیگه رندر بزنم همونو ایمیج2ایمیج میکنه متن نو دیتا عوض میشه مشخصه که تصویر داره میره و میاد ولی نمیدونم چرا از لایه های فتوشاپ نمیخونه هیچی.
ورژن فتوشاپم هم 24
امکانش هست با ورژن ۲۵ به بالا یه تستی بکنید؟@@farhang-n
@@NimaNzrii بله حتما ، خبرشو میدم
اگه لطف کنی و بذاری تو روند حل مشکل کمکت کنم ، کمک بزرگی به پروژه کردی
این ایدی تلگراممه:
NimaNzriii
منتظر پیامتون هستم ❤️
It seems that it is impossible to achieve two diagram guidance, one for IPA and one for controlnet
Anyone know what plugin is being used for the animated noodles?
Hey, great job! Do you think there is a way to connect it while running on Rundiffusion or other cloud services?
I modified the nodes in comfyui, and I want to change the name of each group, or add some groups, but there is no change in photoshop, how can I do this?
following node types were not found:
🔹SendTo Photoshop Plugin
🔹Photoshop ComfyUI Plugin
Sorry i was updating smt , can you update the node and check it again please
May I ask you if using this plugin bypass the photoshop autogenerative rules? EG, they won't generate a Honda or NSFW stuff. Not saying that I need to generate those. But we must not be restricted as Photoshop is just a tool. Not a police to tell me what not to do. And you know what? Why did I ask you that? I can find out myself. lol . Thanks for the video.
You are not restricted;)
I changed the workflow you gave me, and renamed each group in Chinese. When I loaded the workflow again, I couldn't find these groups in the UI, but I could click to generate them. I hope the next version will be improved. Thank you for bringing us such a good plugin.
Can you share
Do we need photoshop 2024 version or does it work with 2023 ?
Nima how can I run this on my laptop as a server with --listen args I connect to my local newtwork from my Apple Mac PC. I use photoshop my mac, and stable diffusion on my windows. It works for web browser comfyui but photoshop it will not work! 4:27 ?
make sure that you dont enter https: before the ip
There is also problem with default instructions with the node titled workflow/math 4 slider (a slider value), node is missing from comfyui manager! work around with right click the existing node and click convert node to widget>convert strength to widget until there is a fix
I got the same error.
When loading the graph, the following node types were not found:
workflow/math 4 Slider ( a slider value )
Nodes that have failed to load will show as red on the graph.
This looks like an awesome tool?! Any chance you’ll be making videos to show us how this runs on the Mac? Thanks for open-sourcing this by the way.
Should be the same process, did you have any issues?
I think 🤔 this node might rival IPA, great work
By the way, I am missing 1 node - Photoshop ComfyUI Plugin.
Please join the discord and send me a screenshot of your problem, we will fix it for you
Does the iPad version have the same plugin support?
Yap
The best comfyUi never!
Thank you for this plugin. Beautiful work.
Please create a tutorial on editing photos in photoshop using comfy - retouching, color grading, relighting
Please
Btw awesome video
Always eager for next videos ❤❤❤❤
Coming soon!
@@NimaNzrii thanks
what version of photoshop does this plugin work with??
PS 2023, Version 25.0.0 or higher
真很棒 加油 (ง •_•)ง
This will be necessary if it allows you to select the drawing area as generative fill
Couldn't understand, can you explain it a little bit?
I usually use Chinese comfyui and Chinese PS, but there will be problems if I don't modify it into English. Chinese comfyui can't recognize parameters such as width, height and seed, and Chinese ps can't recognize selection in comfyui with the local redrawing function. How can I solve this problem
May I know where you downloaded it from? I don't know about Comfi Sue Chinese, please guide me to solve this problem for you
@@NimaNzrii Download this plugin (AIGODLIKE-ComfyUI-Translation) and click switch locale in the bottom right corner, and comfyui becomes Chinese, and then ANYTHING EVERYWHERE in your node's original image about width, height, seeds, etc. That's the first question! The second problem is that the ps language is changed to Chinese, and the version of PS is whatever. It doesn't matter if I use the version of 2024 and 25.10. In fact, when PS is Chinese, no matter whether comfyui is Chinese or English, MASK nodes like #1250 #597 are blank, and the selection of PS cannot be obtained. Therefore, both in-Paint and out paint are invalid IN PS, and everything is normal when ps is in English mode and comfyui is also in English mode. However, such a way makes our foreign users too uncomfortable, I hope you can help solve it, thank you
中文要改下他设置参数的名称 有的要改成中文才能识别
❤
Remote connection not connecting,I use a network gpu server
oh sorry. we are working on network gpus for next ver
Hi, I have some problems and it would be really helpful if you have any clues on how to solve them.
At 6:44 in the Node Photoshop ComfyUI plugin, mine shows "no data."
At 1:24 in Presets, both of the lists show "couldn't load."
Thank you very much!
Can you send me your comfyui terminal l
Log file in discord
@@NimaNzrii hHi, thank you! But the Discord community upside invitation seems to be invalid few days ago.
Sorry , link updated
discord.com/invite/3eHAMWnx7Y
Amazinggggg❤❤❤❤❤🎉 i love u bro
Thanks
Thank You!
Hey! How can I send my pic to layer?
Just click on empty area of plugin
Great work - have already started adding my own sdxl "processes" to the base flow - so much potential. No need to reconnect noodles just send your image to a different process from the plugin
Is this working for sdxl workflow or sd1.5 only?
Can you share your workflow?
Legend
This is epic great work
Wow this tool looks amazing!
What a legend!
Thanks for your brilliant work.
My pleasure Elan🖌️
first one here will try it
Hope you enjoy 🔥
luv your great work!
Thank you! Cheers!
Great work thank you!
Our pleasure!
i just installed the crack version of Photoshop how can I install it?
When you run the node it will automatically install it for you! No matter which ps version you have
Check your ps to see if it worked or not
But if it didn't work just use zxp uxp installer like the video
@@NimaNzrii thanks a lottttt
Adobe 🥱