NVIDIA DeepStream 9.1 Released with New Agent Skills
NVIDIA has released a new version of DeepStream 9.1, featuring 13 agent skills for simplifying video analytics. Among the new features is Multi-View 3D Tracking for tracking objects from multiple cameras.
NVIDIA DeepStream 9.1 Released with New Agent Skills
NVIDIA has released a new version of DeepStream 9.1, which includes 13 agent skills for creating video analytics pipelines.
Instead of manually assembling a computer vision pipeline from scratch, you can now describe the desired outcome in plain natural language. These skills can be used alongside Claude Code or Codex to simplify environment setup, configuration, and execution.
Among the new skills is Multi-View 3D Tracking (MV3DT) for tracking objects using data from multiple cameras, and AutoMagicCalib for automatic camera network calibration.
This release also introduces support for NVIDIA JetPack 7.2 for edge deployment on Jetson Orin and Thor. Everything is published as open-source on GitHub.
Why it matters
AnalysisThis release is significant for developers as it simplifies the process of creating video analytics systems, allowing the use of natural language to describe tasks. The new skills open up new possibilities for automation and improved accuracy in object tracking.
Discuss in community
Share your questions and insights with developers
Related in Ecosystem
Practical guides, core terminology, and agent skills
JSON — Structured Data for Programs, APIs, and AI Agents
A complete hands-on guide to JSON: syntax specifications, serialization via parse and stringify, REST API consumption, JSON Schema validation, and structured outputs for LLMs.
Nvidia's Monopoly and the CUDA Platform
An analysis of Nvidia's technological and economic dominance in the AI market. The CUDA (Compute Unified Device Architecture) platform, created in 2006, transformed ordinary gaming GPUs into the planet's primary computational tool, making it difficult for competitors like AMD and Intel to break this monopoly.
DOCX and XLSX Formats: When AI Needs to Deliver Production-Ready Documents
A comprehensive hands-on guide to generating and processing DOCX documents and XLSX workbooks with AI: comparing against Markdown and CSV, prompt engineering, formula design, and code automation.
TensorRT-LLM & SGLang High-Speed Engines
Deeply compiled computational engines for extreme optimization of language model inference on NVIDIA servers, utilizing graph optimization, FlashAttention-3, and advanced routing.