Getting Started¶

Go from a fresh machine to a running firmware image in minutes. The three guides below cover everything you need — prerequisites, your first build, and a library of ready-made examples to explore.

Install and Setup¶

Install the toolchain, set up the Python environment, and verify everything works with nsx doctor.

Install →

First App¶

Scaffold a project, resolve modules, build the firmware, flash it to an EVB, and stream live SWO output.

Build your first app →

Examples¶

Eight maintained example apps — hello world, CoreMark, PMU profiling, power benchmarking, audio capture, USB serial, and more.

Browse examples →

What You'll Need¶

Requirement	Why
Python 3.10+	NSX CLI and module resolution run on Python
uv	Fast dependency manager used by the project
CMake + Ninja	Build system underneath every generated app
Arm GNU Toolchain	Default cross-compiler for Cortex-M targets
SEGGER J-Link	Flash and SWO viewer (`nsx flash` / `nsx view`)

Optional: armclang (Arm Compiler for Embedded) and ATfE (Arm Toolchain for Embedded) are also supported. See Install and Setup for details.

All of these are covered in detail on the Install and Setup page.

Notation¶

Throughout the docs you'll see one universal placeholder path:

<app-dir> — the directory of a generated or example app (whatever you passed to nsx create-app, or the path to a checked-out example).

Replace it with your actual path when running commands. Contributor-oriented docs may also reference <nsx-repo>, the root of a local clone of the neuralspotx repository — normal app developers do not need one.