deepstream-dev

A coding guide for NVIDIA DeepStream, a toolkit for processing video with GStreamer and running AI models. It covers Python development, TensorRT inference, object detection and tracking, and message-broker integration.

In plain words
What is it for?
Use it when building video analytics pipelines, connecting cameras or video files, adding detection or tracking, or sending results to systems such as Kafka.
Why use it?
It helps avoid incorrect API properties and common setup mistakes by directing development to the relevant reference material.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/nvidia/skills/deepstream-dev
Any agent
npx skills add NVIDIA/skills --skill deepstream-dev
Clone the repo
git clone --depth 1 https://github.com/NVIDIA/skills

Made for: Claude Code, Codex.

Per session 47 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 3,136 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00047 $0.03136
Opus 5 $0.00023 $0.01568
Sonnet 5 $0.00009 $0.00627
Haiku 4.5 $0.00005 $0.00314

Measured 2d ago against content hash 3e42a832ea21, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

deepstream-dev scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/deepstream-dev/SKILL.md · 183 lines

How it starts

The opening of the file, as written. The whole thing — 183 lines — stays where its author put it; the contents beside it link to each section on GitHub.

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 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 when the user requests Kafka/cloud messaging
    • When in doubt, build the minimal working pipeline and let the user ask for additions

Read the full file on GitHub · 183 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 183 lines · 47 tokens per session scan A 3e42a832ea21

Subscribe to this mod's changes

deepstream-dev is a skill published in the GitHub repository NVIDIA/skills (3,144 stars, last pushed 2d ago), licensed Apache-2.0. It adds 47 tokens to every session and 3,136 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

rudder

Use locally captured coding-session intent to resolve a device-local behavioral spec, generate focused tests from that spec, implement the smallest production changes through red-green TDD, and verify coverage with the repository's native tooling. Use when the user asks to run Rudder, create or regenerate tests for…

RudderCode/Rudder · 101 tokens

Monet

Skill "Monet" from Monet-AI-Editor/Monet, covering ai video editor — agent control reference, ⚠️ finding editorctl — mandatory (read this first), ⚠️ output file naming — mandatory (read this first), ⚠️ audio on multi-clip timelines — mandatory and ⚠️ canvas mode — mandatory rules (read this first).

Monet-AI-Editor/Monet · 0 tokens

memoire-design-tooling

Use when a task spans interface understanding, design-system memory, UI audits, design CI, Figma, shadcn or Tailwind code generation, research, or agent design workflows and needs the correct Memi capability selected.

memi-design/memi · 51 tokens

build-swiftui-interface

Use when a coding agent must design, scaffold, implement, or verify an iOS or macOS SwiftUI interface with Apple-platform state, accessibility, availability, testing, and Xcode evidence.

memi-design/memi · 45 tokens

check-changed-folders

Run typecheck, tests, and builds for Rudder's core and plugin workspaces on the current branch versus main, verify Claude/Codex provider parity, verify the centralized agent-instruction layout, and verify agent attribution. Use when asked to run "/check", to validate a branch before commit/PR, whenever a user asks to…

RudderCode/Rudder · 102 tokens

manage-sqlite-migrations

Create and validate Rudder's Drizzle-backed local SQLite migrations. Use when changing src/db/schema.ts, generating or reviewing files under drizzle/, adding or altering database tables, columns, constraints, or indexes, writing data backfills, or troubleshooting migration failures in rudder.db.

RudderCode/Rudder · 62 tokens