Vibe Coding Discovervideo-search-and-summarization
View on GitHubNVIDIA AI Blueprint for video search and summarization (VSS) is a GPU-accelerated reference architecture for building video analytics agents with real-time verified alerts, visual Q&A, and automated reporting. The VSS Blueprint uses vision language models (VLMs) such as NVIDIA Cosmos, LLMs such as NVIDIA Nemotron, RAG, and NVIDIA NIMs.
A GPU-accelerated reference architecture for building video agents that search, summarize, and answer questions about live or recorded footage. Includes MCP-based tools, video RAG, real-time alert verification, and NVIDIA NIM model services.
Use Cases
Search video archives with natural-language queriesSummarize long recordingsAnswer questions about video footageVerify real-time video alertsGenerate video-analysis reportsMonitor smart spaces and warehousesValidate standard operating procedures
Built With
- Language
- Python
- Frameworks
- NVIDIA NIM · Model Context Protocol · RAG · Next.js
Tags
video agents · video search · video summarization · computer vision · multimodal AI · video RAG · MCP · real-time analytics · visual Q&A · alert verification