NVIDIA, ComfyUI Streamline Local AI Video Generation at GDC
Key Facts
- What: NVIDIA and ComfyUI announced App View interface, RTX Video Super Resolution node, and new NVFP4/FP8 model variants for FLUX.2 Klein and LTX-2.3
- When: Announcements made at Game Developers Conference in San Francisco; App View and NVFP4/FP8 models available today, LTX-2.3 NVFP4 support coming soon
- Performance: Up to 2.5x faster performance and 60% lower VRAM usage on GeForce RTX 50 Series GPUs with NVFP4 format
- Tools: RTX Video Super Resolution delivers 4K upscaling 30x faster than popular local upscalers with lower VRAM; available as ComfyUI node and free Python package
- Target: Game developers and creators building cinematic worlds on local RTX GPUs and NVIDIA DGX Spark
NVIDIA and ComfyUI on Monday announced updates that make local AI video generation more accessible for game developers and artists working on RTX-powered systems.
The announcements, made at the Game Developers Conference in San Francisco, include a simplified App View for ComfyUI, integration of NVIDIA RTX Video Super Resolution, and new optimized model variants that significantly boost performance while reducing memory requirements. The updates target creators building cinematic worlds and characters for immersive gaming experiences directly on local hardware.
Simplified Interface Lowers Barrier to Entry
ComfyUI’s new App View presents workflows in a simplified interface for users unfamiliar with node graphs. Artists can enter a prompt, adjust basic parameters and generate content without navigating complex node-based setups. The full Node View remains available, with seamless switching between modes.
The App View is fully compatible with existing RTX optimizations in ComfyUI. According to NVIDIA’s blog post, performance on RTX GPUs has improved 40% since September. With the new NVFP4 and FP8 data formats now supported natively, users on GeForce RTX 50 Series GPUs can achieve 2.5x faster performance and 60% lower VRAM usage using NVFP4, or 1.7x faster performance and 40% lower VRAM with FP8.
New model variants available immediately in ComfyUI include FLUX.2 Klein 4B and FLUX.2 Klein 9B in both NVFP4 and FP8 formats. NVFP4 support for LTX-2.3 is coming soon. Users can download the optimized checkpoints from Hugging Face and load them using ComfyUI’s Template Browser.
RTX Video Super Resolution Brings Fast 4K Upscaling
High-quality 4K video generation has traditionally required balancing speed, memory usage and creative control. Many creators generate lower-resolution previews before upscaling. Traditional upscalers can take minutes to process short clips.
NVIDIA’s RTX Video Super Resolution is now available as a node within ComfyUI, enabling quick 4K upscaling of generated video. The technology runs on RTX GPU Tensor Cores and delivers 4K upscaling 30x faster than popular alternative local upscalers while using significantly less VRAM.
For developers, NVIDIA released a free Python package via the PyPI repository, along with sample code on GitHub and a VFX Python bindings guide. The package is powered by the NVIDIA Video Effects software development kit and provides programmatic access to the same AI upscaling technology used in RTX Video.
Performance Gains on RTX 5090
NVIDIA shared benchmark results for the optimized models on an RTX 5090 GPU. Testing showed substantial improvements for both LTX-2 (512×768 resolution, 100 frames, 20 steps) and FLUX.2 Klein 9B (1024×1024 resolution, 20 steps).
The updates build on NVIDIA’s broader push to bring generative AI workflows to local PCs, giving creators direct control over assets, eliminating cloud costs and enabling faster iteration. This approach has gained traction as developers seek to refine outputs at the pace demanded by real creative projects.
Impact on Game Developers and Creators
The combined updates make professional-grade AI video generation more practical for individual developers and smaller studios working on RTX AI PCs. By reducing VRAM requirements and accelerating both generation and upscaling, the tools allow for more rapid experimentation during concept development and storyboarding phases.
The availability of a simplified App View particularly benefits artists and designers who previously found node-based tools intimidating, potentially expanding the pool of creators who can leverage these technologies in game development pipelines.
What’s Next
NVIDIA will host a training session titled “Create Generative AI Workflow for Design and Visualization in ComfyUI” at GTC in San Jose on March 17. The session, led by NVIDIA 3D workflow specialists, will focus on building RTX-accelerated generative workflows for images, video, 3D and PBR materials.
Additional tools mentioned in the announcement include LTX Desktop, a fully local open-source video editor optimized for NVIDIA GPUs, and LM Link for connecting devices running LM Studio to access remote models.
Developers and creators can get started with ComfyUI through NVIDIA Studio Sessions tutorials, including a guided walkthrough hosted by visual effects artist Max Novak.
Sources
- NVIDIA Blog - ComfyUI Streamlines Local AI Video Generation
- NVIDIA RTX Accelerates 4K AI Video Generation on PC With LTX-2 and ComfyUI Upgrades
- How to Get Started With Visual Generative AI on NVIDIA RTX PCs
All technical specifications, pricing, and benchmark data in this article are sourced directly from official announcements. Competitor comparisons use publicly available data at time of publication. We update our coverage as new information becomes available.