ComfyStudio Pro

@PurzBeats @JulienAIArt cant wait to check this out

Tomorrow @JulienAIArt and I are going to dive into Hermes Agent with the ComfyUI Skill and see what it can do! Bring your questions and tips!

Yeah, I need to try Flux2 Klein 9b. Lots of people have been recommending it to me, and also, it's not as heavy as Qwen Image Edit. Requires less VRAM. I like Qwen ImageEdit; I get good results from it, but I'll try the Flux2. If it's as good as people say, I'll add it to my app sometimes today.

@deadsun_0 @comfystudiopro Only one thing. I don't like the Qwen image edit, even though the Qwen multiple angle workflow it's wonderful I feel Qwen doesn't produce very realistic looking images they always look painterly. I prefer Flux2 Klein 9b

Generating images locally is also an option with ComfyUI. Something I need to try more of. Thanks to @comfystudiopro for showing a demo of character sheet generation.

Yeah, this issue is 100% real. It happens to me all the time. Sometimes you need to install directly from the GitHub repository and not through the manager. Or sometimes you already have it installed and it's still not working. You need to uninstall it and then reinstall it. And on top of that, there might be some custom nodes that are causing conflicts. I agree. It's not easy.

@darklig45208568 @comfystudiopro I just remembered the problem wasn't the ComfyUI app. It was installing the LTX audio-to-video template because there were missing files I had to manually download.

This Grok Imagine template from BMX is ridiculously useful. Save the template, drop in any image, and Grok turns it into a full character sheet in seconds.

@darklig45208568 @deadsun_0 Yeah, I use the ComfyUI portable as well. I think that's the best version to use. Do you think that ComfyUI's interface can be intimidating for new users?

@comfystudiopro @deadsun_0 I use comfyUI portable, very easy to install no hassle. Just uncompres the archive and you are done.

@deadsun_0 your the man! thanks for the shoutout!

Yeah, ComfyUI is a big pain in the ass. I think that's why they rolled out app mode. It's still not easy. The up front cost of a GPU is a lot. I don't know the math/cost of it all but I would think probably using cloud credits is a much more cost-effective way of running generations. By the time you spend what a 5090 cost, there will be a new GPU out and new frontier video models. Yeah I don't know hah. I'm not smart enough to figure that out hah. Seedance is expensive though hah but awesome

@comfystudiopro Comfy was a little difficult to install for me and I almost gave up, but Pinochio was even worse. I do like the idea of local but would probably need to do upgrades if I wanted to do more local. It takes about 12-15 min for a 10 sec video clip in 720p.

Yeah two minutes isn't bad. Though the GPU is quite costly upfront and the electricity is negligible I guess. For testing and experimenting I like local but you really cannot beat the paid frontier models. I'm trying to think of a way: if you're making a music video or a short film, maybe some shots can be run locally and important hero shots can be run on the cloud to save a bit of money but I am not sure. That's why I like using ComfyUI: you can run both. If you're running local or on the cloud, you don't need to change what you're doing. it's all in one app. BUT I get when people say ComfyUI is confusing because I find it confusing as fuck too. This is why I'm trying to make an app to make it easier. For me first, and then if other people like it, that's great.

@comfystudiopro 2 minutes isn't horrible. Thanks for making that demo. I'll try it out. I've really only used LTX in Comfy. Is your goal to generate everything locally?

@deadsun_0 Actually... hah its longer than a minute but not too bad. about 2mins. I go through it here. I'm using my own app for it, but it's the same process through ComfyUI.

@deadsun_0 Just a character sheet takes about a minute with a 5090. One sec I'll show ya.

@TomLikesRobots This is insane

Very interesting. A new multimodal video model from Google apparently called Gemini Omni. (gemini.google.com/share/7d5dc678…) For interest, the one with the lamp showing is SeeDance 2. Gemini does seem to do better at Maths. Prompt: A professor writes out a mathematical proof for trigonometric identities on a traditional chalkboard, explaining the step he is currently on in the equation.

Holllllyyyyyyyy @GeminiApp cooked 😳😳 🚨 Gemini Omni: New video model Here is the first output and see the text coherence , if this is not nano banana moment of video then what is ?? direct link for those who believes otherwise in comments

Music video Workflow in ComfyStudio Pro: Song to Keyframes to Generated Edit. 4k version of music video youtube.com/watch?v=WcHBs-… Here’s the tutorial using ComfyStudio Pro: youtube.com/watch?v=8BsFbU… Download ComfyStudio Pro: comfystudiopro.com github.com/JaimeIsMe/comf… Reddit post with more info reddit.com/r/comfyui/comm… ComfyStudio Pro is free and opensource @ComfyUI @yoland_yan

We’ve temporarily suspended new free-tier account creation on Comfy Cloud due to a recent abuse incident targeting high-cost video generation models. Importantly, - All existing paid accounts and credits are unaffected - Existing free-tier credits will not roll over next month We built the free tier because we genuinely want new users to experience ComfyUI easily. We have to step back temporarily to protect the platform, but we are still working on that goal and will share more updates as soon as we have them. Thanks for the support and patience.

I'm currently training a "DeArchive" IC-LoRA, still cooking, some weird warp in the hair, but for those who'd like to test the early steps with me : huggingface.co/oumoumad/ltx-2…

Within three years @ComfyUI hit 2.5M monthly users and a $500M valuation with zero sales team. Now we're hiring our first Customer Success Manager. Here's why this role is different:

@juliewdesign_ WOW!

further testings for my adaptation of H.G. Wells' "The Country of the Blind" using Nano Banana Pro in Luma (an absolute necessity when building with the creative agent at this project scale) and Seedance 2.0 in Runway

I don't really know what I've made here.