ComfyUI gave me early access and credits to test their new in-app agent, and I've been having fun working with it ahead of today's release! It embodies a lot of what I love about collaborating with AI in a creative process - making technical experimentation faster and more accessible, without sacrificing creative control.
Video made with Comfy Agent + Wan VACE - sound on! 🔊
I've been working with agents + ComfyUI for a while. For me, what makes the new in-app agent special is its ability to work directly inside your workflow, alongside you.
When you send a prompt, the agent sees the current state of your graph, including any changes you made. Its edits appear directly on the canvas, as it makes them (no reloading or opening a new version of the workflow).
It's natural to switch between tweaking parameters and running test generations yourself, and asking the agent to build a new part of the graph or troubleshoot something. It's also good at answering questions to help you get a better sense of what's happening technically, or how various factors relate to your creative intent.
The immediacy of this allows you to quickly test and develop new approaches. The iterative loop is faster, making it easier to experiment freely, and get into a creative flow.
ComfyUI video grid tool, now available as a skill for Hermes Agent (on my Github)
I worked with Hermes to create this skill to compare video tests. Works together with Hermes' built-in ComfyUI integration ✨
more info ↓
@NousResearch@ComfyUI
H3-Tiny shrinks MiniMax H3's video-and-audio model to 17B, about half its size.
It cuts whole blocks with no retraining, and the clips keep their synced sound. A few-step version is on Hugging Face too.
https://t.co/45YGzrRB0u
I like it a lot and it’s clearly very skilled, I can understand it’s not for everyone. It’s crazy to me that people are even calling it “bizarre” or particularly experimental or weird. It’s not even that out there. It’s sad how little exposure many people seem to have to anything outside a really limited set of boring, overproduced pop
cool! that’s pretty fast. There’s also a vcv rack module that creates an mcp server which has been working well for me. i’ve also been creating some vcv modules with coding agents (and then having them patch and test them with the mcp or just by directly creating vcv patches as files)
we went from getting weirdness by default and sometimes trying to tame it, to having to push for and discover weirdness, which i kind of like. i like how it moves the definition of art and craft further in the direction of uniqueness and expression rather than slickness and perfection
Hoy, en el capítulo de "la IA no tiene control", me he visto la keynote de Adobe y he dicho: "ánda, qué curioso... esto lo puedo hacer con mis nodos".
He pulido un par de cosas y he terminado con mi propio método para controlar la acción dentro de un vídeo. Ahora puedo controlar los ritmos, los momentos exactos en los que ocurre algo y un montón de cosas más.
Encima funciona directamente en @ComfyUI sin necesidad de entrenar nada, tanto con Minimax H3 como con LTX. Os muestro algunos test con vídeos y sus curvas de control.
It's so cool to see old techniques - in this case, context smashing by @PurzBeats - resurrected for modern models AND explained nicely
Declining friction for people to understand + use these techniques will lead to such interesting places
Animatediff shall prevail
Context Smashing in H3 is really strange. Using animated masks to create spatially aware context mixing. It averages video/temporal tokens as well as text token blending. Makes for some truly bizarre results.