Wildminder

desktop app for image-to-3D mesh gen via Hunyuan3D-2mini. - Local - open-source - faster shape generation with 8GB VRAM github.com/lightningpixel…

ComfyUI-SegviGen for precise 3D part segmentation. Auto model downloading (Trellis, SegviGen, BiRefNet, DinoV3). Post-processing (VXZ, Latent Slats, Voxel, GLB). github.com/Aero-Ex/ComfyU…

Instruction-guided video editing from Baidu. - type "remove the subtitles" and they instantly vanish. - It flawlessly recreates the hidden moving background. - zero messy blurring or weird pixel patches. - rivals massive paid tools. SAMA easily adds/removes object, changes style without external priors. based on Wan2.1-T2V-14B cynthiazxy123.github.io/SAMA/

Another simple ComfyUI wrapper for Dreamid-Omni. Controllable human-centric audio-video generation with identity preservation. - Talking-head with custom character faces and voices. github.com/benjiyaya/Comf…

Wan 2.1 + LoRA for effective object removing/insertion. EffectErase deals with shadows,reflections and lighting. - no ghosting. - SOTA in temporal smoothness and visual realism henghuiding.com/EffectErase/

converts single 2D images into interactable 3D models; with MonoArt. Interior designers will definitely appreciate this one. - Snap one photo of any furniture - Get a working 3D model in 20 seconds - jointly predicts geometry, part segments, and motion; - built on TRELLIS. Drawers actually slide, doors open.. Test room clearances in VR before buying a thing lihaitian.com/MonoArt/

3DreamBooth- high-fidelity 3D subject-driven video generation. AI just made running an e-commerce brand ridiculously easy. - view-consistent videos from multi-view references; - snap a few photos of your product - turns them into cinematic 3D videos - perfect 360-degree rotation in any scene - zero warped logos or lost textures it's HunyuanVideo-1.5 + LoRA. ko-lani.github.io/3DreamBooth/

some cool stuff from Google! Dropped an Interactive Video Super-Resolution - SparkVSR - CogVideoX1.5-5B; - propagates high-quality textures from edited keyframes across sequences; - tops DOVE and STAR, +24.6% on CLIP-IQA and +21.8% on DOVER; - zero-shot restoration, colorization, style transfer! - beats FlashVSR sparkvsr.github.io

Graphic designers, look away. - Adobe just dropped a new AI model. - It generates fully editable layered files from simple text. - No more flat AI images. - Real RGBA layers with perfect transparency. - It even rips flat JPEGs back into separated layers. - The architecture is honestly beautiful. And... the open-source GitHub link? - no, it's Adobe. Enjoy reading the PDF. arxiv.org/pdf/2603.17965

MooshieUI: Beginner-friendly desktop frontend for ComfyUI. - clean 3-panel layout; - txt2img, img2img, inpainting; - built-in Tiled Diffusion upscaling and MultiDiffusion; - auto-download for models. github.com/Mooshieblob1/M…

3D modeling entirely replaced by stick figures. SK-Adapter brings skeleton-based structural control for native 3D. > Feed it a basic skeleton > Type what you want to see > Get a fully rendered 3D character in under 15 seconds > Already rigged and ready for animation > Zero Blender experience required Game devs are happy. sk-adapter.github.io

POV: you're still using M2.5 while others are happily automating their entire jobs with 2.7 - it writes code to upgrade itself - fixes live server crashes in under 3 minutes - wins gold medals in machine learning competitions - builds full financial models and PPTs from scratch M2.7 is putting in work

SegviGen. Repurposing 3D generative models for part segmentation. Looks cool. Easily select, isolate, or modify any part of a 3D model. -TRELLIS2 with SC-VAE. - tops P3-SAM by 40% fenghora.github.io/SegviGen-Page/

ComfyUI-LCS - steers color directly in the 3D latent subspace during sampling. - 0% structure drift. - No training, no LoRAs. - Model-aware lighting and reflections. FLUX + Z-image github.com/facok/ComfyUI-…

StyleExpert. it's Flux-Kontext with MoE LoRA for semantic-aware image stylization - tops OmniStyle and CSGO; - maps textures and brushstrokes from 1 reference while preserving content. hh-lg.github.io/StyleExpert-Pa…

Tell the old bearded chip designers/ASM programmers they can finally retire (probably). InCoder-32B just dropped. - writes production-grade Verilog - optimizes raw GPU kernels in CUDA - bare-metal C for embedded microcontrollers - handles complete CAD and 3D modeling 32B open-source model for industrial engineering huggingface.co/Multilingual-M…

CS:GO + Adobe + Wan2.1 = WorldCam. Interactive autoregressive 3D gaming worlds. > AI now generates playable 3D worlds live > You move the mouse, the AI builds the map instantly > Turn around, and it remembers exactly what was there No traditional game engines. Just neural networks hallucinating reality. cvlab-kaist.github.io/WorldCam/

new playground from FastVideo. Addictive poison: endless scene tweaks for no reason. Just because it can regenerate the whole video in a second. trippy hallucinations. Session over, I surrender dreamverse.fastvideo.org

Your smartphone camera is about to train your future robotic butler. HSImul3R. Physics-driven 3D human-scene reconstruction. It turns monocular video into simulation-ready 3D scenes by using a physics engine as a supervisor. - contact stability - built on DUSt3R, SAM2, MIDI. yukangcao.github.io/HSImul3R/

Hunter Alpha beats GLM-5? And getting close to Claude Opus 4.6? No way... This mysterious model with 1M context is showing great results with openclaw and jumped into the top 5. claw-eval.github.io