Hardware
Lower the Cost of Building and Running Visual AI Agents with NVIDIA VSS Blueprint 3.3
NVIDIA released VSS Blueprint 3.3, a framework that stitches together vision‑language models (Cosmos), LLMs (Nemotron) and retrieval‑augmented generation to convert live or recorded video into natural
NVIDIA released VSS Blueprint 3.3, a framework that stitches together vision‑language models (Cosmos), LLMs (Nemotron) and retrieval‑augmented generation to convert live or recorded video into natural‑language search, visual Q&A, verified alerts, and automated reporting. The “Build Vision Agent” skill generates multi‑workflow deployments from a single prompt while sharing core infrastructure (Kafka, Redis, Elasticsearch) and can deploy a live bottling‑line solution in under 30 minutes on a two‑GPU RTX PRO 6000 Blackwell host. Adaptive Efficient Video Sampling (EVS) dynamically prunes unchanged visual patches to cut VLM processing latency by 17 % and increase concurrent real‑time streams by 46 %, also enabling 60‑minute video summarization in roughly half the time with 80 % fewer tokens.
Related
- Transform Video Into Instantly Searchable, Actionable Intelligence with AI Agents and Skills
- Develop Physical AI Reasoning, World, and Action Models with NVIDIA Cosmos 3
- NVIDIA JetPack 7.2.1 Adds Agentic Video Skills and T3000 Emulation
- TeleRAG: Efficient Retrieval-Augmented Generation Inference with Lookahead Retrieval
Source: NVIDIA Developer | 2026-09-29