deepstream-dev

Develop NVIDIA DeepStream video analytics pipelines using Python pyservicemaker API.

189|39|Updated May 8, 2026
One-click install
npx skills add https://github.com/NVIDIA/DeepStream --skill deepstream-dev-nvidia
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: deepstream-dev
Source: https://github.com/NVIDIA/DeepStream/tree/main/skills/deepstream-dev
Command: npx skills add https://github.com/NVIDIA/DeepStream --skill deepstream-dev-nvidia

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pyservicemaker, pyyaml, and includes references (resource) components.

What problem does it solve?

This skill simplifies the complex development of NVIDIA DeepStream video analytics pipelines, preventing common configuration errors and architectural pitfalls that often lead to pipeline deadlocks or performance bottlenecks.

Core Features & Use Cases

  • Pipeline Construction: Provides standardized patterns for building GStreamer-based video processing pipelines, including source management, inference, and rendering.
  • Configuration Validation: Ensures correct syntax for nvinfer and tracker configurations, preventing runtime failures.
  • Use Case: A developer needs to build a multi-stream video analytics application that performs object detection and tracking on RTSP camera feeds; this skill provides the exact boilerplate and configuration logic to ensure the pipeline runs efficiently on NVIDIA hardware.

Quick Start

Use the deepstream-dev skill to generate a Python pipeline script that ingests an RTSP stream, performs object detection using a provided ONNX model, and renders the output to the screen.

Frequently Asked Questions about deepstream-dev

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build a GStreamer pipeline for RTSP stream video analytics?

To build a GStreamer pipeline for RTSP video analytics, this skill provides standardized patterns using the Python pyservicemaker API, handling source management, TensorRT inference, and rendering for NVIDIA hardware.

Why does my DeepStream pipeline deadlock or hit performance bottlenecks?

DeepStream pipelines often deadlock due to configuration errors and architectural pitfalls. This skill prevents these issues by enforcing proper memory management, buffer handling, and pipeline state transition requirements.

Can I use pyservicemaker to run video analytics on both dGPU and Jetson platforms?

Yes, pyservicemaker supports developing DeepStream video analytics pipelines for both dGPU and Jetson platforms, ensuring hardware-specific configurations for TensorRT engines and multi-object tracking are correctly applied.

How do I configure nvinfer and tracker elements for object detection in DeepStream?

Configuring nvinfer and tracker elements requires strict syntax validation to prevent runtime failures. This skill validates these configurations and integrates ONNX models for object detection within the GStreamer pipeline.

What is the best way to integrate TensorRT inference engines into a Python video pipeline?

The best way to integrate TensorRT inference engines into a Python video pipeline is using pyservicemaker, which abstracts GStreamer element configuration and ensures production-grade memory management across dGPU and Jetson devices.