Plateforme enterprise pour industrialiser le DeepStream Coding Agent, connecter les providers IA premium, et accelerer les deploiements GitHub, Vercel, Prisma, GCP Marketplace et NVIDIA AI Enterprise.
Le contenu NVIDIA reste intact. La couche Smith-Heffa ajoute une interface operative, une capture de leads enterprise, une console agent live et une trajectoire API / MCP prete pour le go to market.
Un agent specialise NVIDIA DeepStream avec une discipline fullstack enterprise: pipelines video GStreamer, inference TensorRT, tracking multi-objets, Kafka, Kubernetes, GCP Marketplace et interfaces agentiques multi-provider.
nvurisrcbin, nvstreammux, nvinfer, nvosdbin, pipelines production-grade et pattern runtime propres.
TrafficCamNet, YOLO v8-v10-v26, modeles ONNX custom et hot-swap engine file monitor.
Cameras multiples, tracking NvDCF, monitoring FPS, analytics et resilience runtime.
GCP Marketplace, NVIDIA AI Enterprise, Kafka, Kubernetes, Vercel, Prisma et surfaces developpeur.
Prompts rehausses avec une methode PTCF pour rendre les demandes de code plus chirurgicales.
Chat mobile-first, pieces jointes image audio video fichier, et base MCP prete pour la flotte d'agents.
Console conversationnelle mobile-first pour Smith-Heffa Coding. Streaming serveur, routage multi-provider, pieces jointes reelles et base deja prete pour l'extension MCP.
Les prompts DeepStream restent exposes, mais Smith-Heffa peut maintenant les renforcer pour le code avec une structure Persona / Tache / Contexte / Format.
Use DeepStream SDK pyservicemaker APIs to develop the python application that can do the following. - Stream from 4 RTSP cameras simultaneously, decode the videos, batch frames together and infer using ResNet18 TrafficCamNet model. - Use tracker after infer to track the detected
Use DeepStream SDK pyservicemake APIs to develop the python application that can do the following. - Read from files, decode the videos and infer using ResNet18 model. - display the bounding box around detected objects using OSD.
Implement a Python application that uses a multi-modal VLM to summarize video frames and sends summaries to a remote server via Kafka. 1. **DeepStream Pipeline**: Use DeepStream pyservicemaker APIs to receive N RTSP streams, decode video, and convert frames to RGB format. Process
Need to create microservice for the app in @rtvi_app directory. Follow the steps below to complete that. - Create FastAPI based server and implement the endpoints mentioned in the attached image @rtvi_vlm_openapi_spec.png. - Create dockerfile to package the everything together
Use DeepStream SDK pyservicemaker APIs to develop the python application that can do the following. - Read from file, decode the video and infer using ResNet18 TrafficCamNet model. - display the bounding box around detected objects using OSD.
Use DeepStream SDK pyservicemaker APIs to develop the python application that can do the following. - Read from file, decode the video and infer using ResNet18 TrafficCamNet model. - display the bounding box around detected objects using OSD.
Download the YOLO26s detection model using the ultralytics library, then convert the model to ONNX model that supports dynamic batch, in a Python virtual environment. Write a DeepStream custom parsing library for the model. Use DeepStream SDK pyservicemaker APIs to develop the p
Le skill reste le coeur du repo. La couche web le rend presentable, exploitable et prete pour les equipes delivery enterprise.
--- name: deepstream-dev description: NVIDIA DeepStream SDK 9.0 development with Python pyservicemaker API. Use when building video analytics pipelines, GStreamer-based video processing, TensorRT inference integration, object detection/tracking, or Kafka/message broker integratio
--- name: deepstream-dev description: NVIDIA DeepStream SDK 9.0 development with Python pyservicemaker API. Use when building video analytics pipelines, GStreamer-based video processing, TensorRT inference integration, object detection/tracking, or Kafka/message broker integration. license: CC-BY-4.0 AND Apache-2.0 metadata: author: NVIDIA CORPORATION version: 1.0.0 --- # DeepStream Development Skill When this skill is active, **ALWAYS read the relevant reference documents** before generating code. Do NOT rely on memory - the reference documents contain critical details about exact property names, correct API usage, and common pitfalls. ## SDK and Architecture Quick Reference ### DeepStream SDK 9.0 Version Requirements - **GStreamer**: 1.24.2 - **NVIDIA Driver**: 590+ - **CUDA**: 13.1 - **TensorRT**: 10.14.1.48 - **Platforms**: Ubuntu 24.04 (x86_64 and ARM64/Jetson) ### Typical Pipeline Flow ``` Source → Stream Muxer → Inference → [Tracker] → OSD → Renderer ``` Components in `[brackets]` are **optional** -- only add them when the user explicitly requests them. | Stage | Role | Key Element(s) | Required? | |-------|------|-----------------|-----------| | Source | Input from files, RTSP, cameras | `nvurisrcbin` (preferred), `nvmultiurisrcbin`, `filesrc` | Yes | | Stream Muxer | Batches streams for inference | `nvstreammux` | Yes | | Inference | TensorRT model execution | `nvinfer`, `nvinferserver` | Yes | | Tracker | Multi-object tracking across frames | `nvtracker` | **Only if requested** | | OSD | Draws bounding boxes, labels, overlays | `nvosdbin` | Yes (for visualization) | | Renderer | Display or save output | `nveglglessink`, `nv3dsink`, `filesink` | Yes | ### Memory Model DeepStream uses NVIDIA Video Memory Manager (NVMM) for zero-copy GPU buffer transfers. Caps strings use `memory:NVMM` to indicate GPU memory (e.g., `video/x-raw(memory:NVMM), format=NV12`). ## Critical Rules 1. **Only Add Requested Components**: Do NOT add pipeline elements the user did not ask for. - **Tracker (`nvtracker`)**: Only add when the user explicitly requests tracking or object IDs across frames - **Secondary GIEs**: Only add when the user requests classification or attribute extraction - **Analytics (`nvdsanalytics`)**: Only add when the user requests line crossing, ROI counting, etc. - **Message broker (`nvmsgbroker`/`nvmsgconv`)**: Only add whe...
13 references DeepStream structurees pour Codex, Claude Code, Cursor et les equipes delivery qui veulent livrer proprement en environnement enterprise.
This document provides comprehensive best practices, design patterns, and optimization strategies for building production-grade DeepStream applications. These guidelines help ensure performance, reliability, maintainability, and scalability. --- **Best Practice**: Build pipelines
DeepStream Service Maker provides two complementary APIs for custom data injection and extraction: 1. **Media Extractor (BufferProvider/Feeder)** - Inject custom data INTO pipelines 2. **Frame Selector (BufferRetriever/Receiver)** - Extract data FROM pipelines
DeepStream Docker images are hosted on the NVIDIA NGC container registry (`nvcr.io`). They package all SDK dependencies (GStreamer, TensorRT, CUDA, models, sample streams) and require the NVIDIA Container Toolkit (`nvidia-container-toolkit`) for GPU access. - **NGC catalog page**
DeepStream provides a comprehensive set of custom GStreamer plugins optimized for NVIDIA GPUs. These plugins handle video decoding, inference, tracking, visualization, and various other video analytics tasks. Understanding these plugins is crucial for building effective DeepStrea
This document is a comprehensive reference for integrating DeepStream applications with external message brokers. It covers two complementary areas: - **Part 1 -- Kafka Integration Use Cases and Patterns**: Pipeline architectures for streaming analytics data to Apache Kafka, incl
The `pyservicemaker.utils` module provides advanced utilities for extracting frames from media sources with precise control over timing, sampling, and batch processing. These utilities are particularly useful for: - Processing specific time segments (chunks) of video files - Fram
The `nvinfer` GStreamer plugin uses a configuration file to define model parameters, preprocessing settings, and postprocessing options. This document provides a complete reference for all configuration parameters. nvinfer supports **two configuration file formats**: property:
DeepStream supports dynamic addition and removal of video sources at runtime through REST APIs. This capability is built into `nvmultiurisrcbin`, which integrates an HTTP REST server, multiple `nvurisrcbin` instances, and `nvstreammux` into a single GStreamer bin. **CRITICAL: Alw
The DeepStream Service Maker provides a high-level Python API (`pyservicemaker`) for building DeepStream applications. It abstracts away the complexity of GStreamer C API and provides a more intuitive, Pythonic interface for constructing video analytics pipelines. The pyservicema
The `nvtracker` GStreamer plugin provides multi-object tracking capabilities in DeepStream pipelines. It tracks objects detected by inference engines across video frames, assigning unique tracking IDs and maintaining object trajectories. The tracker library (`libnvds_nvmultiobjec
This document provides a quick reference for common errors encountered when developing DeepStream applications, along with their causes and solutions. --- **Symptom**: Pipeline crashes with `RuntimeError: Probe failure` and message `unable to add probe fps-probe`.
This document covers two fundamental DeepStream pipeline construction patterns. **Part 1** explains how to build a simple video player -- reading video from a file or stream, decoding it with hardware acceleration, and displaying it on screen without any AI inference. **Part 2**
The `pyservicemaker` module and its `utils` submodule provide a collection of utility classes for monitoring, configuration management, and helper patterns used in DeepStream application development. This document covers: - **Part 1 -- Performance Monitoring Utilities**: Real-tim
Pipeline de capture de leads enterprise. Chaque demande est ecrite en base Prisma pour qualifier les deploiements go to market.