At this year’s SIGGRAPH conference, NVIDIA is showcasing advancements in AI and graphics research, including new tools for content creation and synthetic video detection. The NVIDIA keynote, held on July 20, highlighted the Model Context Protocol (MCP), which allows AI agents to work within creative applications while maintaining human control over the creative process. The MCP is designed to integrate with various tools, enabling AI to assist with tasks such as texture inspection, color management, and export preparation. NVIDIA said the MCP opens a new chapter in accelerated creativity by making applications agent-ready.
NVIDIA RTX PRO workstations, DGX Spark, and DGX Station systems are optimized to support local AI agents, improving responsiveness and data security. The NVIDIA Agent Toolkit supports MCP integration, offering clients and servers for connecting to remote MCP servers and publishing tools. Adobe, Affinity by Canva, Blender, Boris FX Silhouette, Foundry Griptape, and SideFX are among the companies expanding MCP support, allowing AI assistants to interact with their platforms and automate repetitive tasks. These tools provide AI agents with access to real production context, enhancing efficiency without compromising creative control.
NVIDIA also announced the Synthetic Video Detector, a NVIDIA NIM microservice that helps media workflows identify synthetic content. This tool analyzes video frame by frame to produce a classifier score, aiding editorial teams in prioritizing clips for review or deeper analysis. The tool remains effective through common video processing steps like compression and resizing, supporting trust in video content.
Source: nvidia
