April Wrapped
New models
Seedance 2.0
ByteDance's upgraded video generation model debuted as a partner node on launch day.
Three initial creation methods: convert text, references, or start/end frames into video sequences, offering flexible creativity.
Input AI-produced visuals or real portraits to generate identity-consistent videos with lifelike expressions.
Happy Horse
A multifunctional partner node supporting text/video conversion and editing capabilities.
Wan 2.7
The multi-purpose Wan 2.7 video system is accessible via partner node integration.
Ernie-Image
Baidu’s openly available text-to-visual model using Diffusion Transformer architecture.
Full weight accessibility for customization enthusiasts.
Ace Step 1.5 XL
Commercial-quality open-source music production tools.
Enhanced for cohesive song structures with seamless verse-chorus arrangements.
Quiver
Partner node delivering structured SVG outputs instead of pixel-based images.
Optimal for scalable vector graphics like icons and diagrams without rerendering needs.
GPT Image 2
Available through partner node: combines text-to-image and editing workflows.
SUPIR
Freely available system improving blurry or low-quality images.
Complements upscaling processes.
RIFE & FILM Frame Interpolation
Integrated frame interpolation for fluid motion enhancement or cinematic effects.
SAM 3 & SAM 3.1
Native compatibility with Meta's latest segmentation models.
Veo 3.1 Lite & Veo 3.1 Fast
Partner node now accommodates Google’s efficient Veo variants.
Sonilo
Synchronize video footage with custom soundtracks reflecting motion and emotion.
Feature releases
Execute multiple workflows simultaneously, accelerating batch tasks and experiments.
Significantly improves productivity for daily operations.
ComfyHub is gaining momentum
Nearly 500 shared workflows now available.
Explore existing solutions before building new ones.
Anticipating upcoming innovations!