NVIDIA-AI-Blueprints
- 39 skills
- 0 followers
- 2 days ago last updated
- ▌ Vss Setup Video Analytics API · nvidia-ai-blueprints bundleUse to deploy the vss-video-analytics-api REST service standalone with its Elasticsearch ingest-pipeline, selectable Kafka/Redis stream type, and Kafka-topic readiness gates when Kafka is selected. Not for full warehouse deploy.
- ▌ Vss Deploy Detection Tracking 3d · nvidia-ai-blueprints bundleUse when deploying or operating standalone RTVI-CV-3D / MV3DT multi-camera 3D tracking for calibrated MP4/file inputs and live RTSP streams: missing-calibration handoff to AMC skills, the 4-camera sample dataset, camera config, BEV Fusion, live OSD or saved grid/BEV outputs, bundled brokers, basic external MQTT/Kafka brokers, verification, and teardown. Trigger for generic MV3DT, RTVI-CV-3D, multi-view 3D tracking, multi-cam tracking, or sample MV3DT dataset requests. Explicit warehouse blueprint/profile MV3DT requests route to vss-build-vision-ai; single-camera 2D tracking routes to the 2D tracking or DeepStream skills. Not for full warehouse blueprint deployment, single-camera 2D tracking, camera calibration itself, or VSS summarization, Q&A, and RAG workflows.
- ▌ Vss Benchmark Vlm QA · nvidia-ai-blueprints bundleBenchmark video Q&A accuracy and latency of a deployed RT-VLM (Cosmos Reason 3) via vss vlm run, using questions and videos from the DSS vss-devx-base dataset. Replaces the deprecated nat eval / vss-agent QA path. Not for tool-calling or trajectory evaluation, and not for LVS summarization throughput.
- ▌ Rtvi Vlm Perf Testing · nvidia-ai-blueprints bundlePlan, run, and diagnose reproducible RT-VLM GPU performance canaries and benchmarks. Use this skill when running fresh-container stream-capacity, semantic-isolation, latency, throughput, or regression experiments against an RTVI microservices checkout.
- ▌ Benchmark Video Search · nvidia-ai-blueprints bundleMeasure retrieval quality and latency of a deployed VSS search profile by ingesting a labelled dataset and running the vss CLI across retrieval paths. Use when the user asks to "benchmark video search", "compare search recall between builds", or "profile search latency" on a deployed VSS profile. A default run deletes every uploaded video file on the target deployment before ingesting (RTSP cameras are kept); --skip-ingest deletes nothing.
- ▌ Vss Deploy Warehouse Helm · nvidia-ai-blueprints bundleUse when the user asks to deploy, upgrade, or size the VSS warehouse blueprint (2D / 3D / MV3DT) on Kubernetes via Helm — as opposed to Docker Compose, which is covered by vss-build-vision-ai's warehouse reference. Handles GPU-aware NUM_STREAMS capping so the deployment matches what the perception pipeline can actually sustain.
- ▌ Vss Generate Video Calibration · nvidia-ai-blueprints bundleUse this skill when running or tuning AutoMagicCalib on local MP4s, RTSP, or the bundled sample dataset, or when deploying vss-auto-calibration. Do not use for non-AMC calibration or runtime analytics.
- ▌ Profile Vllm Performance · nvidia-ai-blueprints bundleProfile vLLM and RT-VLM inference to identify and remove GPU-idle gaps, underfilled batches, transfer stalls, serialized multimodal work, scheduler gaps, or KV pressure. Use this skill when GPU utilization is unexpectedly low or bursty, capacity trails another runtime, TTFT or ITL regresses, or multimodal prefill and decode appear serialized. Not for running a first benchmark without a reproducible fixed-load signal.
- ▌ Vss Deploy Video Embedding · nvidia-ai-blueprints bundleUse this skill when deploying, operating, integrating, or customizing the VSS RT-Embed Video Embedding microservice. Covers standalone Docker Compose deployment, the `/v1` REST API for text/video embeddings and live streams, Redis/Kafka/OTel integration, troubleshooting, and bring-your-own-model (BYOM) custom embedding backends, with VideoPrism as an example. Do not use for RT-CV, RT-VLM, VSS Agent, or general VSS deployment work that does not include RT-Embed.
- ▌ Vss Manage Video Io Storage · nvidia-ai-blueprints bundleUse to drive `vss vios` for sensor list, timelines, clips, snapshots, and add/delete of video or stream sources, and the VIOS REST API only for what that CLI does not cover (sensor info/status/settings, storage and recorder status, WebRTC, the RTSP proxy, network scan, device settings, bytes to disk, the NvStreamer API), when the caller names a REST endpoint, or to debug VIOS. Also provisions a source into an Agent-backed or headless deployment, which the deployment fans out to the perception consumers (RT-CV/RT-Embed/RT-VLM). Not for VLM inference, semantic search.
- ▌ Vss Setup Behavior Analytics · nvidia-ai-blueprints bundleUse this skill to deploy the vss-behavior-analytics service standalone (entrypoint, config-source, optional calibration). Not for the full warehouse deploy.
- ▌ Vss Ask Video · nvidia-ai-blueprints bundleUse this skill when answering a question about previously analyzed or freshly scoped VSS video, or when reading a stored VSS memory job or record by id, or whenever a question should be answered by running the `vss memory introspect` command. Route through hot context, agent Markdown notes, `vss memory get` or `vss memory query`, `vss memory introspect`, or an exact-window `vss vlm run`. Not for video retrieval or metadata-answerable questions.
- ▌ Rtvi Byom Porting · nvidia-ai-blueprints bundleUse when adding, debugging, or validating a bring-your-own VLM in VSS RT-VLM, including a custom Hugging Face or NGC checkpoint loaded with configuration only (VLM_MODEL_TO_USE=vllm-compatible), plus adapters, plugins, shims, and model-specific runtime dependencies. Not for ordinary catalog-preset deployment of an already-wired model; use vss-deploy-dense-captioning for that.
- ▌ Vss Search Archive · nvidia-ai-blueprints bundleUse this skill when a user wants to search archived VSS video that is already registered in a configured deployment — by natural-language, similarity, attribute, object-ID, or lexical tag query. Not for fresh clip Q&A, live captioning, video summarization, deployment, or source ingestion/deletion.
- ▌ Vss Query Analytics · nvidia-ai-blueprints bundleUse this skill for read-only incident, occupancy, speed, place, and analytics-sensor questions through the project-local VSS CLI. Not for live VLM, incident-range narrative reports, deployment, or alert-rule management.
- ▌ Vss Summarize Video · nvidia-ai-blueprints bundleUse when summarizing a recorded video through HITL-gated LVS, falling back to `vss vlm run` when LVS is not ready. Not for reports, archive search, or live RTSP captioning.
- ▌ RAG Eval 2 · nvidia-ai-blueprints bundleFilesystem RAG benchmarks: corpus/, train.json, evaluate_rag.py (RAGAS quality). Not for prod monitoring, latency/throughput benchmarking (use rag-perf), or evals outside this repo layout.
- ▌ RAG Perf 2 · nvidia-ai-blueprints bundlePerformance benchmarking for a deployed NVIDIA RAG Blueprint server: profiling pass + aiperf load test driven by a single YAML config. Not for accuracy / RAGAS scoring (use rag-eval) or for deploying / repairing services (use rag-blueprint).
- ▌ RAG Blueprint 2 · nvidia-ai-blueprints bundleNVIDIA RAG Blueprint — deploy, configure, troubleshoot, and manage. Handles any RAG action: deploy, install, start, enable, disable, toggle, change, configure, troubleshoot, debug, fix, shutdown, stop, or tear down any RAG feature or service (Agentic RAG, VLM, guardrails, query rewriting, models, search, ingestion, observability, summarization, reasoning, and more).
- ▌ RAG Eval · nvidia-ai-blueprints bundleFilesystem RAG benchmarks: corpus/, train.json, evaluate_rag.py (RAGAS quality). Not for prod monitoring, latency/throughput benchmarking (use rag-perf), or evals outside this repo layout.
- ▌ RAG Perf · nvidia-ai-blueprints bundlePerformance benchmarking for a deployed NVIDIA RAG Blueprint server: profiling pass + aiperf load test driven by a single YAML config. Not for accuracy / RAGAS scoring (use rag-eval) or for deploying / repairing services (use rag-blueprint).
- ▌ RAG Blueprint · nvidia-ai-blueprints bundleNVIDIA RAG Blueprint — deploy, configure, troubleshoot, and manage. Handles any RAG action: deploy, install, start, enable, disable, toggle, change, configure, troubleshoot, debug, fix, shutdown, stop, or tear down any RAG feature or service (Agentic RAG, VLM, guardrails, query rewriting, models, search, ingestion, observability, summarization, reasoning, and more).
- ▌ Retail Test Runner 2 · nvidia-ai-blueprintsRun the Retail Shopping Assistant test suites under tests/, including offline pytest unit tests in tests/unit and live integration scripts in tests/integration for conversation, quality, timing, and plots.
- ▌ Retail Local Runner 2 · nvidia-ai-blueprintsStart, stop, configure, inspect, and troubleshoot the Retail Shopping Assistant locally with backend Python and React code outside containers, local Milvus infra containers, and remote NIM endpoints from docker-compose-nim-local.yaml.
- ▌ UI 2 · nvidia-ai-blueprintsDiscoverability shim for the Retail-Agentic-Commerce UI Codex skill. Use when creating or modifying frontend pages, components, hooks, styling, tests, or browser behavior.
- ▌ Setup 2 · nvidia-ai-blueprintsDiscoverability shim for the Retail-Agentic-Commerce setup Codex skill. Use when the user asks to setup, install, start, launch, or run the project stack.
- ▌ Features 2 · nvidia-ai-blueprintsDiscoverability shim for the Retail-Agentic-Commerce backend feature Codex skill. Use when writing or modifying Python backend code, routes, services, models, tests, or protocol logic.
- ▌ Cve Remediation 2 · nvidia-ai-blueprintsDiscoverability shim for the Retail-Agentic-Commerce CVE remediation Codex skill. Use for CVE, BDSA, GHSA, OSV, Dependabot, Black Duck, Snyk, npm, PyPI, or package advisory remediation.
- ▌ Pre Commit Analysis 2 · nvidia-ai-blueprintsDiscoverability shim for the Retail-Agentic-Commerce pre-commit analysis Codex skill. Use before commits, after significant changes, or for implementation-quality validation.
- ▌ Retail Test Runner · nvidia-ai-blueprints bundleRun the Retail Shopping Assistant test suites under tests/, including offline pytest unit tests in tests/unit and live integration scripts in tests/integration for conversation, quality, timing, and plots.
- ▌ Retail Local Runner · nvidia-ai-blueprints bundleStart, stop, configure, inspect, and troubleshoot the Retail Shopping Assistant locally with backend Python and React code outside containers, local Milvus infra containers, and remote NIM endpoints from docker-compose-nim-local.yaml.
- ▌ UI · nvidia-ai-blueprintsReact and Next.js frontend development standards for Retail-Agentic-Commerce. Use when creating or modifying UI pages, components, hooks, styling, tests, or Apps SDK widget browser behavior in src/ui or related frontend surfaces.
- ▌ Setup · nvidia-ai-blueprintsLaunch all services for the Retail-Agentic-Commerce project. Use when the user types setup, install, or asks to start, launch, or run the project stack.
- ▌ Features · nvidia-ai-blueprintsPython backend development standards for FastAPI, SQLModel, pytest, Ruff, and Pyright. Use when writing or modifying backend code in src/merchant, src/payment, src/apps_sdk, agents, tests, API routes, services, models, or protocol logic.
- ▌ Cve Remediation · nvidia-ai-blueprintsDependency vulnerability remediation workflow for this repository. Use when fixing CVE, BDSA, GHSA, OSV, npm, PyPI, Dependabot, Black Duck, Snyk, or other package vulnerability findings, especially when the user asks for minimal package-only updates, validation, commits, branches, or pull requests.
- ▌ Pre Commit Analysis · nvidia-ai-blueprintsPre-commit implementation analysis for Retail-Agentic-Commerce. Use before commits, after significant changes, or when validating Apps SDK tools, mock data, MCP communication, widget isolation, or implementation quality.
- ▌ Deploy · nvidia-ai-blueprints bundleDeploy Nemotron Voice Agent via root compose using recipe profiles. Use when deploying or troubleshooting auth/startup.
- ▌ Upgrade Pipecat · nvidia-ai-blueprints bundleUpgrade the Nemotron Voice Agent to a new Pipecat (pipecat-ai) version. Reads release notes for every release in range, diffs old vs new, discovers every example pipeline and Pipecat call site, implements changes, then runs multi-agent gap analysis until clean. Generic across Pipecat versions.
- ▌ Configure Pipeline · nvidia-ai-blueprints bundleConfigure Nemotron Voice Agent runtime via `.env`, example-local `services.{cloud,local}.yaml`, and example-local `prompts.yaml`. Use when changing prompts, tracing, audio knobs, exposed pipelines or transports, or local NIM image overrides.