EASY Inpainting in ComfyUI with SAM (segment Anything) | Creative Workflow Tutorial

Outpainting please

what would be the best way to use this to correct badly generated hands? When I try to use this workflow, it's still generating bad hands, even if the masking and everything else is working perfectly xD

Great tutorial, simple and easy to understand, thank you so much!

good clip, thanks, it is funny the load image node is not to be found under loaders if you dont have that node search plugin installed...

i played around with successfully with clipsegs txt2mask, your approach looks much more interesting.

The whole point of comfy is to be able to reproduce workflows by simply importing a json or photo w/ metadata. So tired of seeing tutorials and guides that don't have the file attached.

hi thank you soo much for creating this awesome tutorial is there any way to tell that it generate more than one resault ? (like Empty Latent Image node )

also, music is great!

never mind I found it didn't know it was built in to comfyui, to change this just goto settings at the bottom under link render mode select straight

how did you get your Node Lines to be striaght like that as you create the connections? thanks

Nice and simple, Thanks. Subscribed. Could you explain how to save this workflow so i can pull it out everytime is needed, please? And any other simple additional workflows such as Controlnet, Image to image etc? Thanks and keep them coming

What a great video!

awesome ! can you share your Json worfklow file ?

Please for all that is holy, the new and old gods. please make a tutorial in how to create custom addons/nodes. I searching and did not yet found good docs.

Great stuff, I saw an alternate way where you can use a word instead of dots, so "T-shirt" to get the SAM to select the T-shirt and change it's colour.
Has anyone asked yet...What are you using to make the AI generated voice-over....it's very good, but not perfect 😉
(don't be saying....damn! that my real voice dude!)😁

Is there a way to do inpaint with Comfyui using Automatic1111's technique in which it allows you to apply a resolution only to the mask and not to the whole image to improve the quality of the result?

К сожалению, это несколько не то, чего бы мне хотелось. Как сделать так, чтобы именно дорисовать. Например, у меня есть арт, где нарисован ковер. И я хочу, чтобы с помощью inpaint можно было выделить область, прописать туда "человека сидящий на ковре" и нейросеть нарисовала его в том же стиле, что и основной арт. Пока у меня не получается достичь результатов, которые выходили, когда я пользовался Автоматиком.

This looks great! I will come back when I am ready to fully embrace ComfyUI. Invoke has nodes now so I'm uncommitted.. And thank you for NOT adding music to your videos while the instruction is happening.

i installed the comfyUI impact pack but the SAm detector does not seem to be working. Is there any fixes you would suggest for this?

You know in A1111, the inpaint is such that only a portion of image is inpainted while nothing is done to rest of the image but in comfy The whole image is rendered even though only a small part is changed.

Is it possible to do A1111 like inpaint in comfy?

This is going to be so useful. Thank 😊 you!

Great video, thanks