ComfyUI v0.43.0 Adds New Nodes and Model Features

What happened

ComfyUI v0.43.0 includes synchronization of graph cleanups for model_prefetch, contributed by rattus128. This update also adds rope scaling support to MOSS-image, contributed by kijai. Partner-node sign-in tokens were renewed in API nodes by dante01yoon. HeyGen Voice instant clone and text-to-speech functionality were added via Partner Nodes by bigcat88.

The Crystal Video Upscaler node was added via Partner Nodes, also by bigcat88. FLUX 3 Video nodes now support 2K/4K resolutions, contributed by bigcat88. MOSS-image color-depth control was added by kijai. The Readme was updated to mention MOSS image support by comfyanonymous. FLUX 3 Max nodes were added via Partner Nodes by bigcat88.

Why it matters

These updates expand ComfyUI's capabilities in video and image generation, making it more versatile for developers. The enhancements to Partner Nodes improve integration with external services, offering more tools for content creation. The addition of new models and features supports a broader range of creative workflows.

Key details

The release includes contributions from multiple developers, including rattus128, kijai, dante01yoon, and bigcat88. Version v0.43.0 incorporates changes such as FLUX 3 Video nodes supporting 2K/4K resolutions and the Crystal Video Upscaler node. Workflow templates were updated to version v0.11.92, ensuring compatibility with new features.

The top-p sampling threshold was corrected by curealb. The Readme was updated by comfyanonymous to reflect support for MOSS image. These details highlight the collaborative nature of the release and its focus on improving user experience and system stability.