I want confirm in short, it works with:
rocky linux flame ←→ comfyui on rocky linux with nvidia (local & lan server)
rocky linux flame ←→ comfyui on apple silicon (local & lan server)
rocky linux flame ←→ comfyui on windows with nvidia (local & lan server)
apple silicon flame ←→ comfyui on rocky linux with nvidia (local & lan server)
apple silicon flame ←→ comfyui on apple silicon (local & lan server)
apple silicon flame ←→ comfyui on on windows with nvidia (local & lan server)
My initial h.265 frustration got me hired to become an enterprise ml/ai software integrator recently.
I will dedicate my spare time for the community of course.
If you need a proper h.265 / hevc import from 12/10bit footage from iphone , pro & consumer cameras for your rocky linux flame, I welcome you to give me a pm to test it out.
I am not sure anyone realized The GitHub as a platform of worldwide software engineers playground is on stake. GitHub now is Microsoft CoreAi. Change? Subtile day by day. This looks like the End of OpenSource / Public Domain. Bill closed this Gate… remotely? On purpose? Nobody know about that? Nobody ? And Nobody care? I am speechless.
end of an era…
Everyday I get new ideas to fool around with. You know I made some pretty complex workflows and developed some nodes. Now I have the super easy task to only translate subtitles as srt into any other language. reliable. fully automated. Yes, I am on it. And I want it to translate all 2678 lines of the srt. Not only 62
Not so much flame related anymore, I did develop a simple but powerful translator for davinci subtitles with inhouse power. PM me if interested
I mean local translation with ollama. It works crazy cool with the right setup. No Data ever go out of your LAN System.
After nearly 3 years using ComfyUI by my self, I think it is the most adapted framework for generative AI with a broad open source support today. Meanwhile I can say it did not give bread’n butter for a daily life; programming nodes and integration solutions for a small surviving niche such as VFX. To be honest, I rarely touched comfyui within the last 3 months. It’s pretty cool, but I am working now 99% more with node.js as a javascript based framework, and not so much python anymore. I dropped Flowise as well for being useful, but headed to n8n as the most sophisticated AI platform recently. Today I use very much Open-WebUI together with embedded ComfyUI and N8N chains for generatives and daily life tasks. Personally I see the world focus shifted from python based solutions to node.js based integrations. I am not sure what comes next. Please let me know how you think!
@Xteve - good work on your flame integration and for this, python is the only path.
as for node, the truly remarkable thing is the silence of the planet after realizing that npm was so compromised, while simultaneously being the primary loading mechanism for gemini.
the mind boggles and the meat jiggles.
yes I know npm got about a 30 minutes compromise on some critical packets couple of weeks ago. Still this is the most active community today and I can review any code that I propably get to use. I love python because it’s simple and easy. But I also like node.js because I am able to make it redundant running without ever thinking about: Is my server down? In fact, revisiting any code before use is absolutely recommended. Bit nostalgic : time flies so fast. today a single average generalist developer (like me) can develop in a day or two in a street cafe; that was impossible 1999 or 2000 with 6x top expert java developers in a month with 12x bug tracing testing personnel on former high end sun solaris machines. interesting times indeed.
@Xteve - All very true and the coding models get more robust every day.
The first iteration of a sidewalk cafe, vibe coded project may not always be 100% robust or complete, but by that measure, it can also be argued that large mature codebases are equally incomplete or insecure.
Personally, I’m being trained by machine learning to be a better coder.
cool man I think so. Do not trust any “help” from models. Review it. Sometimes they blow up the code until the models halluzinate. Thats the point of trashing. I like coding agents in ASK mode only. so I can peer review BEFORE i commit anything. We are so old school. But it still works best ![]()
I am still here watching and trying to fix anythig if there is any request
Please just stop this nonsense already.
yes, I agree. But what nonsense? Let me know please. I will stop any nonsense immediately if I know what the nonsense is
I believe what my esteemed colleague is trying to say is please keep this to Autodesk Flame and Flame-related discussions.
Thanks.
EXR 16bit /32 bit is working pretty well in different color spaces with Flame and ComfyUI. Single Layer or Multi Layer (Multi Channel) supported. If someone need a custom implementation, please let me know.
I’m interested in giving this a try. is it still in development? Is there a demo that shots exactly how It works? thanks in advance.
Hi AdamF,
ComfyUI is an ever evolving system made by and for creatives. Still the IO possibilities are a catch, since genZ user base only ask for png and some other formats far from Industry Standard. That’s why I did build a bunch of ComfyUI Custom nodes for Audio/Video Industry Standard Input and Output FileFormat’s that is able to read and write pictures as single layer(channel)EXR as well unlimited multilayer (multichannel) EXR from and to Flame (C4D, Blender, Houdini..you name it). For Example send a multi channel EXR sequence from flame to comfy… Do whatever you want in Comfyui on ANY or EVERY of those possible layers…send it back as multichannel EXR back into flame… and process it further inside flame. The interested user base was (within the last year) just a VERY limited community, … mostly top end VFX studios around the world. Personally I love EXR as image container as well 24bit Audio WAV. Well, I got hired as AI Implementor for automotive automation pipelines in June last year in Asia… so my available support window is now weekends via google meet or zoom or teams or wechat. Let me know what you really need and lets figure a chat with a life demo
Best, Xteve

