ComfyUI is the most flexible, composable, and powerful open-source media generation tool with a massive ecosystem of workflows and custom nodes. Your Hermes Agent can now install, launch, manage, and run sophisticated ComfyUI workflows on demand.

@ComfyUI LFG

@ComfyUI holy shit

@ComfyUI why do you keep tantalizing us with these sound bites and leave me still needing an album...or playlist.

@ComfyUI It's coming!

@ComfyUI i am just the voice of the streets...and kudos to you all for shipping like crazy

@ComfyUI Daaaamn! Every day it's something new. Bookmarked. This will save time. Next we need Unreal Engine automation.

@ComfyUI Just in time!

@ComfyUI stop it. stop making my custom workflows so easy. its creeping me out. first azure foundry. now comfy. add demucs. add outlook. add teams. add azure. then basically all the infra I setup you integrated lol

@ComfyUI Hermes it’s just on another Level.

@ComfyUI Things are getting outta control quick! 🔥

@ComfyUI damn that slaps

@ComfyUI Hey @3duaun did you see this? Your dreams are coming true!

@ComfyUI Let's go! ComfyUI is tough! But not anymore!

@ComfyUI

Okay, this is weird. I vibe coded with Claude same thing, initially as skill but then rewrote as plugin. Here is the problem - there is native image gen tool, adding image gen skill confuses agents. I hit reliability issue. So I rewrote it into plugin. Big advantage is that there is no decision tree for agent, agent can't get confused, agent uses native image gen tool (which registers as provider), agent don't have to parse any JSONs and, finally, plugin can spit out MEDIA tag that can be picked up natively, low res preview can be sent back to the model to use native vision (another plugin). You already have image gen plugins, why not extend it with comfyui one - that sounds very logical.

@macbethAI @ComfyUI I've been using it without an official skill. How do I install this official skill?

@macbethAI @ComfyUI Just update Hermes or point it at this

@ComfyUI @PurzBeats @JulienAIArt

@ComfyUI i... i was building this for the hackathon... WHY?!

I do use @ComfyUI almost daily, but is on a different machine as I believe is the case of most people (that do inference already borderline, living in the edge of VRAM limits) so I hope this is taken in account. Would be absolutely awesome doing the main operations on the main inference machine and calling for image/video crafted by the designer machine in which Comfy lives.

@ComfyUI Constantly cooking

@ComfyUI Hermes literally runs better everyday. Totally local. Gets everything done. Except for running a 𝕏 bot. It’s not very good at that. But I don’t think that’s what’s it’s for. But great work.

@ComfyUI Lovin' this addition.

@ComfyUI oh me oh my

@ComfyUI Hermes is on fire, I love it!

@ComfyUI Beautiful

@ComfyUI

@ComfyUI @kaiotei_ try this as soon as you can, just take whatever you use atm & drop it into this, you'll see within hours whats there in the architecture

@ComfyUI ayooooooooo lesgooo!!!!

@dillonrolnick @ComfyUI

@ComfyUI do you guys really have these many specialist? you guys churn out good integrations overnight

@ComfyUI Comfy uwu

@ComfyUI when you say "now" do you mean this is a new feature in a new release or a PR, or its been there.. Basically, what does one do to activate it, because this working can confuse me. Thanks for clarifying in advance.

@ComfyUI It was just merged, you can update Hermes to get it

@ComfyUI I remember joking about AI taking peoples jobs...

@ComfyUI 这太棒了 可以使用开源模型批量生产图片了 只是llm的token有点贵啊 我怕养不起Hermes 你们有好办法吗 本来使用comfyui本地工作就是为了节约成本的

@Teknium @ComfyUI Wow are you serious?