Skip to main content

Installation

Install the Client SDK to control and interact with Aria Gen 2 devices from your PC.

Prerequisites​

  • Hardware: Aria Gen 2 device
  • Mobile App: Aria Companion App installed and paired with your device (setup guide)
  • Connection: USB cable for device-to-PC connection

Supported Platforms​

OS / PlatformOS / Distro DetailsAria Gen2 Support (projectaria-tools ≥2.3.0)Supported Python Versions
Linux (x64)Ubuntu 22.04 LTS (jammy) / 24.04 LTS (Noble Numbat), and recent Fedora releases. Requires glibc 2.34 or newer✅ Supported3.10 – 3.12
Linux (ARM64 / aarch64)e.g. NVIDIA Jetson. Same glibc floor as x64🧪 Experimental — partial toolchain3.10 – 3.12
macOS (Apple Silicon / ARM64)macOS 14+ (Sonoma or newer) on M1/M2/M3/M4✅ Supported3.10 – 3.12
macOS (Intel / x64)—❌ Not supported—
Windows (x64)—🚧 Planned—
Linux ARM64 is experimental, and only part of the toolchain runs there

projectaria-client-sdk and projectaria-tools publish aarch64 wheels, so the SDK installs on ARM64 Linux hosts such as NVIDIA Jetson. Not everything follows: projectaria-vrs-health-check has no ARM64 wheel, so VRS Health Check, MPS and Aria Studio cannot be installed there. Treat ARM64 as usable for device control and capture, not as a drop-in replacement for an x64 host.

No macOS Intel builds

No x86_64 macOS wheels are published, so pip install fails outright on Intel Macs rather than installing an older build. Apple Silicon is required.

Ubuntu 22.04 Known Issue

If you encounter a GLIBCXX_3.4.31 not found error when running CLI commands on Ubuntu 22.04, you'll need to update your C++ standard library. See the GLIBCXX troubleshooting guide for the solution.

BLE control has a narrower platform matrix

The table above covers the SDK as a whole. BLE device control is experimental and is not supported on Ubuntu. See BLE Device Control.


Step 1: Install Python​

To use the Client SDK, you'll need Python 3.10 - 3.12.

Step 2: Create a Virtual Environment​

rm -rf $HOME/projectaria_gen2_python_env
python3 -m venv $HOME/projectaria_gen2_python_env
source $HOME/projectaria_gen2_python_env/bin/activate

Step 3: Install the Client SDK Python Package​

pip install projectaria-client-sdk

pip pulls the runtime dependencies in for you — you do not install them separately:

DependencyVersionUsed for
projectaria-tools~=2.3.0VRS reading, device calibration, sensor data types
projectaria-login~=0.2Device authentication
rerun-sdk~=0.33.0The aria_streaming_viewer visualization window
numpy, termcoloranyArray handling and CLI output
Rerun is pinned to a minor version

rerun-sdk is pre-1.0 and changes its Python API across minor releases, so the Client SDK pins the minor it is built against. If your environment already has a different rerun-sdk, install into a fresh virtual environment (Step 2) rather than forcing the versions together — a mismatched Rerun fails when aria_streaming_viewer starts.

Verify Installation​

After installation, run aria_doctor to configure your system and verify the SDK is working:

aria_doctor

Expected behavior: The command opens required ports on your PC for device discovery and configures your system to allow unblocked internet connection.


Available CLI Controls​

The Client SDK provides a CLI to control your Aria Gen 2 device:

  • Recording: Start, stop, list, and download recordings in VRS format
  • Streaming: Start and stop streaming sensor data
  • Device: Obtain device information, connect to Wi-Fi, send text-to-speech, and more

See the CLI Quick Reference for a command overview or the Technical Specs for the full reference.

Next Steps​

  1. Authenticate your device to enable SDK access
  2. Explore Recording to capture data
  3. Check out Streaming to stream data in real-time