ComfyUI ecosystem updates — open-source model releases, custom nodes, workflows, and tools for image, video, and audio generation.
Text to videoTr1dae ships a trained clean-latent 2x upscaler for MiniMax H3 with ComfyUI nodes, keeping audio intact between samplers. Community tests: 0.5MP to 1080p in under 5 minutes.
Video editingComfyUI H3 Motion Context MultiRef Update 6 rebuilds long-form MiniMax H3 workflows: a 20-slot music video chain and an AV extension chain with seamless audio.
Text to videoDownload the MiniMax H3 Spatial Physics LoRA for ComfyUI: improves object collisions, stacking, falling, and gravity-consistent motion in text-to-video generation.
Text to imageSenseTime's official ComfyUI node pack (v0.2.0) runs SenseNova-U1.5 locally: native 4K text-to-image, single/multi-image editing and region control at about 17GB peak VRAM.
Text to videoA ComfyUI timeline editor for MiniMax H3: drag images, video and music onto tracks, write a prompt per shot, and preview the exact storyboard prompt before rendering.
Video editingDownload the LTX-2.5 CQ enhancer LoRAs: two rank-128 adapters (image + video) that restore and enhance old, low-resolution, poor-quality footage in ComfyUI, with bundled workflows.
Text to videoAlaya Lab releases EVOKE, a 14B open-weights world model generating 384x640 video with 3 steps, external world-state memory, and mid-flight prompt steering.
Apache-2.0
MultimodalMiniMax releases Music 3, an open music model creating complete songs up to 5 minutes with expressive vocals, structured captions and lyrics, with native ComfyUI support.
Text to videoByteDance releases Bernini-Diffusers-v2: the full semantic-planning video generation and editing pipeline in diffusers format, with stronger reference-guided editing and OpenS2V.
Apache-2.0
ComfyUI v0.33.1 brings native MiniMax Music 3 support with core CUDA Graphs, new MiniMax and Bria partner nodes, Anima tunes with extra blocks, and several stability fixes.
Image to videoDownload MiniMax H3 Turbo Ref2V v0.1: minimax_h3_ref2v_turbo_4step_v0.1_comfyui_bf16.safetensors, a 4-step Ref2VA LoRA for multi-reference video, ready for models/loras.
Text to imageGazingstars123 releases Anima-2.9B preview v1, a community fine-tune expanding Anima from 28 to 40 layers, trained on 1.7M anime samples with a dedicated ComfyUI node.
circlestone-labs-non-commercial-license
Image to videoTencent ARC releases SCoPE, adding sightline-coordinate positional encoding to Wan2.2-I2V-A14B so videos follow a camera trajectory while keeping the image-to-video prior.
Apache-2.0
EcosystemComfy MCP local server: install comfy-mcp with pip, then run your own ComfyUI from Claude Code, Claude Desktop and Cursor. Introspect nodes, run workflows, route by hardware.
Text to videoInner-Reflections releases a MiniMax H3 LoRA for looping anime-style sketches: rough textured outlines, flat minimal coloring and seamless loop videos with native audio in ComfyUI.
Text to videoDownload MiniMax H3 Hybrid: FL2VA quality plus Ref2VA references, including minimax_h3_hybrid_fl2va_ref2va_b25-49-int8.safetensors, with the ComfyUI Hybrid Loader node.
Text to videoA new ComfyUI node suite stitches FL2VA keyframe clips into long MiniMax H3 videos, preserving motion and native audio across segment boundaries with example workflows.
ComfyUI v0.32.0 adds a built-in attention backend matching SageAttention speed, native LTX 2.5 support with STG and dual CFG, MiniMax H3 VAE and memory fixes, plus partner nodes.
Text to videoDownload LTX-2.5 for ComfyUI: Lightricks 22B open audio-video model with native multishot, 4K HDR, official T2V/I2V workflow templates, int8 and NVFP4 quantized checkpoints.
Image to videoDownload MiniMax H3 Turbo LoRA v1.0: minimax_h3_fl2v_turbo_8step_v1.0_comfyui_bf16.safetensors and the 4-step 768p ComfyUI variant, ready for models/loras.