Voice Command — Cortex Forge App

We've been building on NVIDIA Jetson since the TX1 in 2015 — over a dozen projects across TX1, TX2, Xavier, Orin, and Thor. This is our attempt to open-source some of the frameworks and building blocks we use internally to kickstart every new project. Sharing what we know helps us grow too. Hope it does the same for you.

Overview

A standalone client application built on the Cortex Forge platform. Part of a suite of 10 apps that demonstrate real-world ML inference workflows on NVIDIA Jetson AGX Orin.

Directory Structure

CortexForge/
├── core/
│   ├── cortex-forge-driver/       # Linux kernel module
│   ├── cortex-forge-server/       # gRPC server
│   └── cortex-forge-gui/          # Operator console
│
└── apps/
    └── cortex-forge-app-voice-command/          # This app

Quick Start

cd apps/cortex-forge-app-voice-command
make deps
make proto
make run SERVER=localhost:50051

gRPC API Dependencies

Method	Purpose
LoadModel	Load a model onto the platform
Infer	Single-shot inference
InferStream	Continuous streaming inference
GetMetrics	Pull live telemetry
HealthCheck	Verify server connectivity

Author

Sandesh Ghimire — sandesh@soccentric.com

Copyright (c) 2026 SoC Centric LLC. All rights reserved.