hashu786/comfyui-tessera is a collection of custom nodes that upscales video inside ComfyUI. The latest release is v1.0.0. The repository description presents the use case of taking 544p footage to 4K or 8K on a single consumer GPU, based on MiniMax H3. However, the information provided does not separately list anything newly added or changed in v1.0.0.
The processing flow divides the video into spatial tiles and time segments and streams the encoder and decoder. It describes a configuration that does not load all full-resolution pixels into memory at once. The node collection includes tile planning, guide anchors, fine-tuning controls, and latent representation storage for recovery. The README also organizes the installation procedure, example workflows, and criteria for choosing settings. The implementation language is Python and the license is MIT. The purpose of this configuration is to split and process video upscaling work within ComfyUI and to reduce memory usage.
Figures for performance and memory usage cannot be confirmed from the README description alone. It is therefore difficult to judge, from the information presented, the environments in which the 4K·8K output and the single consumer GPU condition actually hold. Tiling and time-segment processing and the streaming configuration are the key elements, but concrete performance evidence remains unverified.