Joined February 2017
158 Photos and videos
This one was really fun to create together with @Stefan_3D_AI . We collected the most crucial tips to level up your 3D AI Workflow. It just launched on the channel, hope you will enjoy it as much as we did!
Most people who hate on 3D AI just do not know how to build a proper workflow around it. That is why it starts to feel like the tools are still too rough and not ready for production. So my colleague @philippsieben and I put together one video covering the main problems and what actually helps solve them. Watch the full handbook here πŸ‘‡ youtu.be/xRUhx8Ft5bQ #3DAI #Blender3D #ComfyUI #GameDev
2
47
Triposplat is out and its insane! This is probably the easiest local 3D AI setup I've ever shown, no installs needed & open-source πŸ”₯ TripoSplat is natively built inside ComfyUI and can be used on only 6gb vram. βœ… 3D Gaussian splats from a single image βœ… Zero dependencies or custom nodes βœ… Works on Windows, Linux, Mac AMD βœ… Exports directly as mesh to Blender βœ… MIT license β€” free commercially Full setup guide πŸ‘‡ youtu.be/Ldwd1APMw5A
1
3
229
The New Open-Source 3D AI Generator Pixal3D just dropped and I show you how you can run it locally on only 6gb vram on your own! Pixal3D has: βœ… Pixel-perfect front-view accuracy βœ… Runs on just 6GB VRAM βœ… MIT license β€” free for commercial use βœ… GGUF quantized version in ComfyUI βœ… Better mesh quality than Trellis 2 Full setup guide here πŸ‘‡ youtu.be/LMmuhIwaeB4 #3DAI #Trellis2 #ComfyUI #OpenSource #Pixal3D πŸ”₯
7
198
πŸ”₯ Free vs Paid 3D AI in 2026 β€” who actually wins? I tested Trellis 2 (free local) vs Tripo AI (paid cloud) across 4 real assets 🎯 βœ… Stylized character βœ… Sci-fi hard surface prop βœ… Architectural building βœ… Organic plant Then took the winner through full retopo, rig and animation 🎬 Go check it out here πŸ‘‡ youtu.be/CzY_qDPip4Q

3
208
🚨 The first local 3D AI Studio is here and it's FREE πŸ”₯ 3DGen Studio gives you a real studio interface for all your open-source 3D AI workflows 🎨 βœ… Trellis 2, Hunyuan3D, Qwen Image β€” all in one place βœ… Full Kanban pipeline πŸ“‹ βœ… Mesh editor texture inpainting πŸ–ŒοΈ βœ… Graph mode for branching workflows πŸ”— βœ… 100% local, 100% free Full setup guide on the channel πŸ‘‡πŸŽ¬ youtu.be/acNNQsIgejs #3DAI #ComfyUI #Trellis2 #OpenSource #LocalAI πŸ”₯ Video: ug.link/dxp4800plus-rin-aT6I…
3
387
This is incredible news for the 3D AI Space πŸŽ‰ Can't wait to cover Pixal3D in a dedicated video soon.
‼️UPDATE: Pixal3D is now under the MIT License. We hope this makes Pixal3D easier to use, build upon, and adopt for broader research and applications. Thanks again for your support, and feel free to let us know your feedback!
1
154
My Trellis2 GGUF workflow video hit 100k views in 2 weeks 🀯 A lot of you have been asking for the full setup, workflow files, the exact node config, model links, settings cheat sheet, plus the Blender cleanup pipeline I didn't have time to show in the video. So I'm packaging it all up: Trellis2 Workflow Pack β€” launching in ~6 weeks. Includes: β€’ The hero GGUF Low-VRAM workflow (runs on 6GB cards) β€’ Full precision workflow for 16GB GPUs β€’ Step-by-step PDF guide (setup β†’ first asset β†’ cleanup β†’ export) β€’ RunPod template for Mac users / no-GPU setups β€’ Prompt library settings cheat sheet πŸ‘‰ Waitlist gets early-bird pricing first access: pixel-artistry.com/Trellis2W… What would you most want me to cover as well? Drop a comment πŸ‘‡
2
176
Most people don't know NVIDIA quietly dropped a free Text-to-Motion AI Kimodo runs fully offline on just 6GB of VRAM and generates 3D animations from text prompts. πŸ”₯ I created a detailed video about it: ⚑ Install Kimodo locally 🎯 Generate animations from text prompts 🦴 Rig your character with Mixamo 🎬 Retarget everything in Blender with the free Rokoko add-on Zero cost. Zero cloud. Just your GPU. πŸ’» Install scripts and workflows here πŸ‘‡ pixel-artistry.com/Kimodo (youtu.be/nNmeuvew8LY) #3DAI #Kimodo #Blender #AIAnimation #ComfyUI
2
203
Most people think you need a high-end GPU to run Trellis 2 locally. You don't. GGUF quantization lets you run the best open-source image-to-3D model on just 6GB of VRAM β€” and it's actually faster than the full model. Here's exactly how to set it up in ComfyUI Q8 = fastest Q4 = lowest VRAM Both beat BF16 on speed AND memory Quality difference? Basically invisible. Watch the full video here youtu.be/FuFm8zBHDWI #Trellis2 #3DAI #ComfyUI #AITools #3DModeling
3
216
PixelArtistry retweeted

142
60
641
449,844
PixelArtistry retweeted
New 3D AI News episode #8 β€” in case you missed it πŸ‘€ #NeuralFur is the most interesting here: it goes beyond static fur generation and reacts to physics and edits. That makes it feel much closer to something usable in a real 3D workflow. Full breakdown with my colleague @philippsieben πŸ‘‡ youtu.be/DGn6EM_8Rts #3DAINews #3DAI #ComfyUI #OpenSource3D #Generative3D
2
1
52
3,275
PixelArtistry retweeted
3D AI News #8 is where open-source 3D gets interesting again. 🚨My colleague @philippsieben and I put together a latet 3D AI news roundup. What’s inside: 🐾 #Neural_Fur for fur and groom 🧱 Pascal Editor for walls, buildings, and props πŸ›  #Hitem3DSkills #MeshyMCP #RodinMCP βš™ ComfyUI 3D Viewer Pro #Kimodo 🌊 See-through #Seedance3.0 Featuring @ComfyUI , @MeshyAI , @_ArcadeStudio_ and @wildmindai . Full episode πŸ‘‡ youtu.be/DGn6EM_8Rts #3DAINews #3DAI #ComfyUI #OpenSource3D #Generative3D
1
10
925
πŸ‘€
AI Native Seams... Finally? #3D πŸ”₯ Excited to share our recent work MeshTailor: learned from real artist seam layouts, it places edge-aligned, coherent seams on low-poly meshes in seconds, enabling strong UV unwrapping. πŸ”— Check out the paper below! ⬇️
1
3
189
Dedicated ComfyUI tutorial about it, will launch this weekend on my channel πŸ™ŒπŸ»
Thanks for developing our SeviGen into @ComfyUI ! SegviGen, the SOTA 3D part segmentation method, has been conditionally accepted by SIGGRAPH 2026 (Journal Track). Feel free to try it! Project: fenghora.github.io/SegviGen-… Code: github.com/Nelipot-Lee/Segvi…
1
3
224
PixelArtistry retweeted
Replying to @VARCO3D
@VARCO3D is looking wild! The next step in 3D gen we were looking for. I shall try it out today <3 Prompt Mesh Edit. Texture generation and texture painting Propper remesh for lowpoly game assets Check out @philippsieben video if you need convincing youtube.com/watch?v=IaEXLG4L…
1
1
4
236
Open-Source 3D AI Local Gen is just getting started! Check out in my new video how you can generate highly detailed 3D AI Models locally, entirely for free with Trellis2 FaithC. The setup is really easy and workflows are free for you to download! youtu.be/R25nUevmXRU?si=vDT4…
1
3
228
Local 3D AI Generation is actually getting good! I used Sam3 for masking, Trellis2 for 3D Model Gen and SegviGen for part-segmentation. All local, open-source and in less than two minutes' time. Check out more of my tutorials here: youtube.com/@PixelArtistry_
2
3
8
324
Very great to hear that!
I sent a message out to the ComfyUI community on stability and what we are doing about it: reddit.com/r/comfyui/s/gxdgS… TL;DR: Comfy has grown fast in users, contributors, and complexity. While trying to keep up with all use cases, platforms, and new models, we let velocity outrun stability. However, Comfy is infrastructure for a lot of people's workflows, experiments, artistic expression, and even livlihoods. Regressions are not just annoying anymore -- they break things people depend on. We want Comfy to be something you can rely on, and are shifting everything towards working on stability and consistency in the coming months. In the post, I discuss some concrete steps and processes. Thanks to everyone for supporting us and holding us to a high bar!
1
118
I just shared in my most recent video how you can use Hunyuan3D 3.1 in ComfyUI via API and how you can use the new globally released Hy3D Studio to generate, retopologize, and animate your 3D AI Models. You can check it out here: youtu.be/mD1pG01ndnc?si=8_lr…
1
3
227