| name | inference-builder |
| description | Generate deployable Vision AI pipelines with high-performance video and
streaming capabilities on NVIDIA GPUs. Activate when users want to: create
GPU-accelerated inference microservices or standalone apps for vision,
video, or streaming workloads; write or edit pipeline YAML configs; build
Docker images for GPU inference; work with models from NGC or HuggingFace;
or deploy with DeepStream, Triton, vLLM, TensorRT-LLM, or PyTorch backends.
|
| owner | chunlinl@nvidia.com |
| service | inference |
| version | 1.0.0 |
| reviewed | 2026-04-09 |
| author | NVIDIA |
| tags | ["inference","nvidia","deepstream","vllm","tensorrt-llm","triton","docker","gpu"] |
NVIDIA Inference Builder
A code generator for Vision AI pipelines with high-performance video and
streaming capabilities on NVIDIA GPUs. It turns a YAML config (+ optional
OpenAPI spec and custom processors) into a deployable inference pipeline,
packaged as a Docker container or standalone application.
Setup
A .skill_config file in the same directory as this SKILL.md provides the
project root and ready-to-use commands. Read it first — it contains:
INFERENCE_BUILDER_ROOT — absolute path to the project repositoryINFERENCE_BUILDER_VENV — command to activate the virtual environmentINFERENCE_BUILDER_CLI — command to invoke the builder CLI
Example .skill_config:
INFERENCE_BUILDER_ROOT="/home/user/inference-builder"
INFERENCE_BUILDER_VENV="source /home/user/inference-builder/.venv/bin/activate"
INFERENCE_BUILDER_CLI="python /home/user/inference-builder/builder/main.py"
To run the builder from any working directory:
source /path/to/.skill_config
eval "$INFERENCE_BUILDER_VENV"
$INFERENCE_BUILDER_CLI --help
If .skill_config does not exist (e.g. running directly from the repo), all
paths below are relative to the repository root and the CLI is:
source .venv/bin/activate
python builder/main.py <config.yaml> [options]
First-time venv setup
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
Getting current CLI usage
Always run with --help for authoritative flags — do NOT rely on memorized
syntax:
python builder/main.py --help
Key Directories
