NVIDIA-AI-BlueprintsOfficial·22 Agent Skills Included

video-search-and-summarization

Deploy and operate GPU-powered video search, summarization, and alert pipelines

Deploys NVIDIA's Video Search and Summarization stack and operates its microservices for video Q&A, summarization, semantic archive search, and real-time alerts. Removes the manual work of wiring Docker Compose profiles, NGC model pulls, camera onboarding, and Elasticsearch analytics queries. Guides AI agents through readiness checks, incident reports, and calibration so video workflows run end to end with fewer errors.

1,839|384

↗

npx skills add NVIDIA-AI-Blueprints/video-search-and-summarization --all -g -y

Available:

Tells the AI agent how to route user requests to the correct VSS skill, run deployments and docker commands itself in the background, query the analytics server with the required session handshake, and report public access URLs instead of internal ports.

All Skills in This Repository (22)

Pure Emerald Level Indicators

📦 In Repo

NVIDIA-AI-Blueprints

vss-frag

Generate video summary reports using Long Video Summarization and Milvus knowledge retrieval.

vios

Manage video sensors and streams via VIOS REST API calls.

deploy

Automate deployment, validation, and teardown of NVIDIA VSS profiles with Docker Compose.

video-analytics

Retrieve video incidents, alerts, and metrics from Elasticsearch.

rt-vlm

Deploys the VLM service for real-time video captioning and incident detection.

alerts

Manage and monitor video alerts via the VSS agent's natural-language interface.

video-understanding

Analyze video frames to answer visual content questions via the VSS agent.

video-search

Search stored videos by natural language queries for objects, actions, or events.

report

Generate timestamped captioned video summaries via VSS REST API queries.

vss-deploy-profile

Deploy, verify, debug, and tear down NVIDIA VSS Docker Compose profiles on GPU hosts.

vss-deploy-detection-tracking-2d

Deploy, debug, and operate the RTVI-CV 2D detection and tracking microservice via Docker and REST API.

vss-summarize-video

Summarize recorded video clips via the LVS microservice with a VLM fallback.

Official

Advanced

Frequently Asked Questions

FAQPage Schema

How to install video-search-and-summarization?▼

Run `npx skills add NVIDIA-AI-Blueprints/video-search-and-summarization --all -g -y` in your terminal to install all skills in this suite globally.

What does the NVIDIA VSS blueprint do?▼

It lets you search video archives with natural language, summarize long recordings, ask visual questions about clips, and verify real-time alerts using vision-language models on NVIDIA GPUs.

How do I deploy a VSS profile with an AI agent?▼

Ask your agent to deploy a profile such as base, search, lvs, or alerts; the vss-deploy-profile skill handles credentials, compose dry-runs, GPU sizing, and readiness checks automatically.

Can I search recorded video with plain English?▼

Yes. The vss-search-archive skill ingests video files or RTSP streams, builds embeddings, and returns timestamped clips matching queries like 'find all forklifts'.

Do these skills work with Claude Code and other AI agents?▼

Yes. Every skill follows the standard SKILL.md format and has been evaluated with Claude Code and Codex, so they run in any compatible agent environment.