# video-search-and-summarization | NVIDIA AI Blueprints | Knowledge Base | Instagit

Suite of reference architectures for building GPU-accelerated vision agents and AI-powered video analytics applications.

GitHub Stars: 942

Repository: https://github.com/NVIDIA-AI-Blueprints/video-search-and-summarization

---

## Articles

### [How the ChartGenerator Creates Visualizations from Video Analytics Data in NVIDIA VSS](/NVIDIA-AI-Blueprints/video-search-and-summarization/chart-generator-video-analytics-visualizations)

Learn how ChartGenerator creates visualizations from video analytics data It converts JSON results into charts uploads them and returns image tags for easy UI rendering

- Tags: how-to-guide
- Published: 2026-05-15

### [Understanding the video_frame_timestamp Tool and Its Timestamp Mapping in VSS](/NVIDIA-AI-Blueprints/video-search-and-summarization/video-frame-timestamp-tool-usage)

Explore the video_frame_timestamp tool and its precise timestamp mapping. Extract video frames and get accurate UTC times with this powerful NVIDIA AI tool.

- Tags: deep-dive
- Published: 2026-05-15

### [How to Debug VSS Deployment Issues Using the Deploy Skill and Docker Compose Logs](/NVIDIA-AI-Blueprints/video-search-and-summarization/debug-vss-deployment-issues)

Debug VSS deployment issues effectively. Use the deploy skill for container health checks, pipeline validation, and inspect Docker Compose logs with resolved.yml for quick troubleshooting.

- Tags: how-to-guide
- Published: 2026-05-15

### [How the RTVI VLM Alert Tool Processes Video Streams: A Technical Deep Dive](/NVIDIA-AI-Blueprints/video-search-and-summarization/rtvi-vlm-alert-tool-stream-processing)

Explore how the RTVI VLM alert tool processes video streams discover its real-time captioning and anomaly alert mechanisms via SSE and Kafka.

- Tags: deep-dive
- Published: 2026-05-15

### [VSS Incident Formatting: How the Multi-Incident Formatter Workflow Works](/NVIDIA-AI-Blueprints/video-search-and-summarization/vss-incident-formatting-workflow)

Understand VSS incident formatting. Learn how the NVIDIA multi_incident_formatter workflow normalizes timestamps and generates JSON payloads with video URLs and charts.

- Tags: how-to-guide
- Published: 2026-05-15

### [How to Integrate Custom VLMs with video_caption and video_understanding Tools](/NVIDIA-AI-Blueprints/video-search-and-summarization/integrate-custom-vlms-vss-tools)

Effortlessly integrate custom VLMs with video_caption and video_understanding tools using Nat Builder. Register compatible wrappers without altering core code. Enhance your video analysis workflow now.

- Tags: how-to-guide
- Published: 2026-05-15

### [VSS VST Tool Implementations for Timeline, Sensor List, and Duration Coverage](/NVIDIA-AI-Blueprints/video-search-and-summarization/vss-vst-tool-implementations)

Explore VSS VST tool implementations for timeline, sensor list, and duration. Query VST for temporal boundaries, sensor discovery, and video durations using LangChain.

- Tags: architecture
- Published: 2026-05-15

### [How the attribute_search Tool Works with Elasticsearch Query Builders in NVIDIA VSS](/NVIDIA-AI-Blueprints/video-search-and-summarization/attribute-search-tool-elasticsearch-query-builders)

Explore how the attribute_search tool uses Elasticsearch query builders for efficient video analysis. Discover its two-stage query pipeline for KNN vector search and result refinement.

- Tags: how-to-guide
- Published: 2026-05-15

### [VSS code_executor Docker Backend Implementation for Sandboxed Execution](/NVIDIA-AI-Blueprints/video-search-and-summarization/vss-code-executor-docker-backend)

Implement secure VSS code_executor Docker backend for sandboxed Python code execution. Discover strict security policies, non-root containers, and resource limits for isolated tasks in the Video Search & Summarization platform.

- Tags: how-to-guide
- Published: 2026-05-15

### [How VSS Evaluators Use LLM-Judge Functionality to Score Trajectory QA Report Quality](/NVIDIA-AI-Blueprints/video-search-and-summarization/vss-evaluators-llm-judge-usage)

Discover how VSS evaluators use LLM-judge to automatically score trajectory QA reports. Learn about single-field comparisons and dynamic discovery for quality assessment.

- Tags: how-to-guide
- Published: 2026-05-15

### [How VSS Handles Video Upload and RTSP Stream Ingestion via FastAPI](/NVIDIA-AI-Blueprints/video-search-and-summarization/vss-video-upload-rtsp-ingestion-fastapi)

Learn how VSS handles video upload and RTSP stream ingestion using FastAPI. Discover its efficient streaming PUT and RESTful endpoints for seamless data processing and embedding generation.

- Tags: how-to-guide
- Published: 2026-05-15

### [Architecture of the multi_report_agent for Video Analysis Reports: NVIDIA VSS Agent Design](/NVIDIA-AI-Blueprints/video-search-and-summarization/multi-report-agent-architecture)

Explore the multi report agent architecture for streaming video analysis and unified markdown reports. Understand the NVIDIA VSS Agent design and its three layer NAT framework.

- Tags: architecture
- Published: 2026-05-15

### [How to Extend the VSS Agent with Custom NAT Tools and Function Registration](/NVIDIA-AI-Blueprints/video-search-and-summarization/extend-vss-agent-custom-nat-tools)

Extend the NVIDIA AI Toolkit NAT agent by registering custom tools and functions. Learn how to create Python modules, import them, and expose them to the framework.

- Tags: how-to-guide
- Published: 2026-05-15

### [Security Considerations for Deploying VSS with NVIDIA NIM: Best Practices and Checklist](/NVIDIA-AI-Blueprints/video-search-and-summarization/vss-nvidia-nim-deployment-security)

Discover essential security considerations for deploying VSS with NVIDIA NIM. Learn best practices for API key isolation network configuration and GPU memory limits to secure your deployment.

- Tags: best-practices
- Published: 2026-05-15

### [How the embed_search Tool Performs Multi-Embedding Fusion for Video Archives](/NVIDIA-AI-Blueprints/video-search-and-summarization/embed-search-tool-multi-embedding-fusion)

Discover how embed_search fuses multiple embeddings for video archives. It uses text, image, or video inputs, Elasticsearch KNN, and fusion strategies for optimal retrieval.

- Tags: deep-dive
- Published: 2026-05-15

### [Best Practices for Real-Time Video Alerts with VLM Verification in NVIDIA VSS](/NVIDIA-AI-Blueprints/video-search-and-summarization/vss-real-time-alerts-vlm-verification-best-practices)

Implement real-time video alerts with VLM verification using NVIDIA VSS. Discover best practices for automatic incident detection and efficient stream processing. Learn to leverage the POST /v1/generate_captions_alerts endpoint...

- Tags: best-practices
- Published: 2026-05-15

### [How NAT Tools in VSS Retrieve Video Clips and Snapshots from Video Storage Toolkit (VST)](/NVIDIA-AI-Blueprints/video-search-and-summarization/vss-nat-tools-vst-interaction)

Learn how VSS NAT tools simplify video clip and snapshot retrieval from VST. Discover typed interfaces, retry logic, and URL rewriting for efficient access to video data.

- Tags: how-to-guide
- Published: 2026-05-15

### [Long Video Summarization Chunking and Aggregation: How the LVS Microservice Processes Long-Form Video](/NVIDIA-AI-Blueprints/video-search-and-summarization/lvs-microservice-chunking-aggregation)

Learn how the LVS microservice processes long videos by chunking, aggregating, and deduplicating timestamps for a concise, unified summary. Explore the NVIDIA AI Blueprints video search and summarization repo.

- Tags: how-to-guide
- Published: 2026-05-15

### [What Is the Model Context Protocol (MCP) in VSS for Video Analytics?](/NVIDIA-AI-Blueprints/video-search-and-summarization/vss-model-context-protocol-mcp)

Discover the Model Context Protocol (MCP) in NVIDIA VSS for video analytics. Learn how MCP exposes video data as callable tools for LLMs.

- Tags: internals
- Published: 2026-05-15

### [How to Configure Video Analytics Elasticsearch Integration in VSS](/NVIDIA-AI-Blueprints/video-search-and-summarization/configure-vss-elasticsearch-integration)

Easily configure video analytics Elasticsearch integration in VSS. Set environment variables and use ESClient for asynchronous queries to incidents and frame data. Get started today.

- Tags: how-to-guide
- Published: 2026-05-15

### [Understanding VSS Deployment Profiles: dev-profile-base, dev-profile-search, dev-profile-alerts, and dev-profile-lvs](/NVIDIA-AI-Blueprints/video-search-and-summarization/vss-deployment-profiles-explained)

Explore VSS deployment profiles like dev-profile-base, dev-profile-search, dev-profile-alerts, and dev-profile-lvs. Package distinct AI capabilities from basic video understanding to live summarization and incident detection.

- Tags: architecture
- Published: 2026-05-15

### [How the VSS Agent Architecture Integrates with NVIDIA NIM Microservices](/NVIDIA-AI-Blueprints/video-search-and-summarization/how-vss-agent-integrates-nvidia-nim)

Discover how the VSS agent architecture integrates with NVIDIA NIM microservices via service discovery, OpenAI-compatible endpoints, and LangChain abstractions. Learn more!

- Tags: architecture
- Published: 2026-05-15