Refiner ✦

BLOOD made them BROTHERS... 🩸🩸 a THRONE turned them into ENEMIES. 👑⚔️ Witness the tragic tale of two Yeti brothers. ❤️‍🔥❤️‍🔥 THE LAST OF THE YETIs Made with @YouArtStudio using Seedance 2

@Spectromachina "Tank, load the jump program!"

you can learn anything with ai, literally just open any software you want to learn and start doing stuff, just ask ai questions and it will answer, like in the matrix when they learn kung fu instantly, never thought id be making game assets, let alone at this speed!

@PhotogenicWeekE @BUTA_no_rakuen Hou much GPU and RAM for this quality?

@BUTA_no_rakuen はい、完全ローカルです。ComfyUI + LM Studio(Prompt解釈)。LTX-2.3、頑張れば何とかここまでは(笑)

Fable 5が使える間にwと、Seedance 2.0用Prompt、どこまでLTX-2.3 t2vで行けるか？実装してみた。基本timelineの解釈だけど、いろいろ(勝手に)ディレクションして飽きない様工夫を。Prompt一文字も変えてません。どうでしょうか？(笑) 顔の一貫性はLoRA使用

@JayRetro85 @grok @imagine Terrific shot!!! 🔥🔥⚡⚡

I don’t like having to rotate my phone but liking these shots. Do you prefer vertical or wide? Experimenting with wide shots for the Jay Retro series. Made with @grok @imagine

@StevieMac03 @nikitabier Hello, please check your DM.

Stolen works every day it seems, some raking in millions while my original posts my get in the tens of thousands. @nikitabier it would be great to know if the new system picked these up. This one never even bothered to remove the watermark... x.com/GothicOrnate/s…

@joaoalexmachado @Primee32 Thank you 🙂‍↕️

@Refiner_Studio @Primee32 Maestro of robbery 😂

THE 16-FRAME GRID TRICK THAT KEEPS SEEDANCE 2.0 CHARACTERS CONSISTENT One monk. One white dragon. A full chase sequence through a waterfall canyon and into a forest. No drift, no dragon losing its horns mid-scene, no robe changing shade between cuts. Most creators generate one reference image then re-prompt Seedance 2.0 shot by shot and hope the character holds together. The real consistency comes from building the entire sequence as one 4x4 storyboard grid in GPT Image 2.0 first — then feeding that grid into Seedance 2.0, not raw prompts. Here's the workflow: 1. Write the character bible — monk: bald, gold robe. dragon: white fur, antler horns, amber eyes. locked before frame one 2. Draft 16 frame captions — shot type + action + one punchy line each 3. Lock the visual style in a single block — cinematic, shallow depth, water mist, motion blur on fur 4. Generate the full 4x4 grid in GPT Image 2.0 — numbered corners, thin black borders, captions baked in 5. Pull each frame out as its own reference image 6. Feed each frame + its caption into Seedance 2.0 as an individual shot 7. Stitch the shots in storyboard order Why this works: - Dragon and monk get designed once across the whole grid — not reinvented in 16 separate prompts - Motion blur and mist already encode energy so Seedance has less to guess - Frame captions double as your shot list — no re-planning once you're animating - Consistent framing forces consistent camera logic before video even starts Use cases: ⁃ Mythology / fantasy sequences with non-human creatures ⁃ Action chase scenes across multiple environments ⁃ Character + creature pairs that need to stay locked across cuts ⁃ Cinematic shorts with no budget and no team Not every fur detail survived the jump to video untouched — some cleanup needed once things start moving. But designing the dragon and monk together in one grid before touching Seedance killed almost all my consistency issues.

@IamEmily2050 @Primee32 There's a saying, "speaking with confidence makes even the Petty seem Strong!" Thank you for making me aware 🙏

@Refiner_Studio @Primee32 It is not his work and he have no idea what he is talking about 🤣 because he did not generate the video or wrote the prompt 🤣🤣

@IamEmily2050 @Primee32 Oh 🤦🤦🤦

@Camcamera11 Any link or source where I can get references for prompts that work better with LTX?

@Refiner_Studio Because prompts that work well for WAN aren't necessarily the best for LTX. Lately, I've been feeding my LLM an LTX-specific prompting guide, so it already knows how to structure prompts 😉

I sat down to edit a video. Walked away with a blooper reel instead. No notes. This is the better version anyway😆 LTX 2.3 DaVinci Resolve 16GB VRAM

@Camcamera11 @krea_ai @ComfyUI Thank you 🙂‍↕️

@Refiner_Studio @krea_ai @ComfyUI Really like the colors here 😊

My first attempt with @krea_ai 2 using @ComfyUI This is just a test, my first try with local Ai generation. Not a paid promotion. (Krea2 Darkbrush safetensor)

@DannySwift29036 @openart_ai @WonderStudiosX 🔥🔥👏👏

This is one of my productions with open art director. @openart_ai @wonderStudiosX #WonderOADirector

@jtydhr88 This is beautiful work! We'll need to refine our workflows with our own tools.🪄💫⚡

A feature I’ve always wanted to add to ComfyUI’s 3D nodes — camera presets. It’s still a WIP; I even built a dedicated camera path editor just for it, and open-sourced it. But I’m not sure when it’ll finally make it into the native codebase, or even whether it can be added at all… Still, at least I’ve taken this first step. If it’s possible, I’d also like to bring in a pose editor, what’s now commonly called a “3D director’s console”, but is that even feasible? I honestly don’t know…

@IamEmily2050 @michaelricks @hellorob Thank you so much 🙏🙂‍↕️🙇

@Refiner_Studio @michaelricks There are so many Node but always use the official ComfyUI node or someone you trust, follow ComfyUI people and check their YouTube channels,@hellorob check Rob account for more indep information and always try the open source models improve your skills before using API credit.

Here's what I see~ we're being fed a load of BS on the compute requirements. Don't believe me? Download LTX 2.3 with ComfyUI and test it yourself. Unlimited FREE generations. How fast? Depends on your machine. I'm using a 5-year-old RTX 3090 with 4 slots maxed out with 64gb total memory. Video generates in about 1-2 minutes, 5 seconds, 720p. Krea 2 much faster. Beautiful image generation. LTX ComfyUI integration uses approximately 28-30% of my memory during generation. Krea 2 only 12%. Is LTX 2.3 equivalent to Seedance 2.0? Not yet, but the model is amazing as it is. This is true UNLIMITED. And UNCENSORED. The control is in your hands, not theirs. The future is OPEN SOURCE and local. All these companies charging by credits have built their business model on a House of Sand, imo. @ltx_io @ltx_model @ComfyUI

@IamEmily2050 @michaelricks What would be the best way to integrate Seedance 2 (as API) in my Comfy UI workflow (recently received some decent credits to check the API version? Or is it better to just stick to browser generating for Seedance 2?

@Refiner_Studio @michaelricks Far away from Seedance V2 but good to do alot of things, hopefully LTX 2.5 will be close to Seedance V2.

Thanks for sharing this. I have just recently started tinkering with comfyui (got a new machine 2 days ago: Ryzen 9, 64gb RAM (6000mts), rtx 5060ti 16GB) I would appreciate if you can please 🙏 help me with a few questions: Does it give good control: "directing" a scene? Prompt adherence? Video/camera references? Audio? Also, how much graphic memory you have in RTX 3090? Any links to configure this easily and start using fast?

@IamEmily2050 @michaelricks Does it give good control: "directing" a scene? Prompt adherence? Video/camera references? Audio?

@michaelricks Hopefully LTX 2.5 this month and there should be another open source Chinese model also coming this month 🤞

@vladimircherner Thanks for sharing this 🙂‍↕️

Higgsfield has just wiped out all the competition!! Using its supercomputer capabilities, one creator has developed an app that mimics Cinema 4D. Watch the video to see exactly how it works. The cool thing is that you can set any camera movement (any!) and recreate any actions. Then run it through the Seedance 2.0 renderer, and your video – with maximum control – is ready. This is the very same functionality that’s usually built using the Unreal Engine in professional productions. And here, it’s available for the price of a subscription. The app is available here higgsfield.ai/supercomputer/…. Just a heads-up: this is only suitable for pros. It’s just cheaper)

@jphilipp @taherdhanera Very well said! Making a movie and promoting a movie to reach a bigger audience are two different tasks requiring different types of expertise.

@taherdhanera YouTube could be that, with marketing done through trailers shared on X etc. Nice thing you can then also watch on TV through the likes of Apple TV, because really it can become a couch experience like Netflix. But it's tough -- not every filmmaker is a marketer, too.

Serious AI films deserve a serious platform. Right now, social media rewards 15 second dopamine clips. Random visuals get millions of views, people celebrate the numbers, and suddenly everyone calls themselves an AI filmmaker. But filmmaking has never been about chasing viral moments. It is about storytelling, emotion, craft, and years of learning. AI is a powerful tool, but it cannot replace the DNA that real filmmakers build over time. A viral AI clip does not make someone a filmmaker. It just means the algorithm liked that clip. AI filmmaking is a serious craft. It deserves a space where people watch stories, not just scroll past them.