Neat Library

Use this page to install or update the Palette Neat Library. You can install it directly on a DevKit, or inside the Neat Development Environment when you build from a host machine.

The Neat Development Environment includes a Neat Library version, but the Neat Library can also be installed or updated independently. That means you can install a newer compatible library without reinstalling the SDK container. For supported combinations, see the Compatibility Guide.

Choose Where to Install

Run the installer from the environment where you build or run Neat applications:

Environment	Use this when
Neat Development Environment with DevKit Sync	You want the host build environment and paired DevKit runtime kept in sync automatically.
Neat Development Environment without DevKit Sync	You only need to build in the container and will transfer binaries to a DevKit manually.
DevKit	You want to install or update the runtime directly on the device.

DevKit Sync keeps the runtime aligned

When you install from inside an SDK container paired with a DevKit through DevKit Sync, the installer also installs the Neat Library on the paired DevKit. This keeps the host build environment and the DevKit runtime on the same Neat Library version.

Prerequisites

Install sima-cli in the environment where you plan to install the Neat Library. Use the sima-cli installation guide.

Before you begin:

• Make sure the environment running the installer can download SiMa.ai artifacts.
• We recommend creating an account at community.sima.ai to access assets such as precompiled models for testing.

Prepare an Install Directory

Choose a working directory before running the installer. On the DevKit, install under /media/nvme to avoid filling the root file system. In the SDK container, use a container-local path such as $HOME/neat.

sdk-container$cd $HOME && mkdir -p neat && cd neat

Install the Default Package

Use this option to install the default Neat Library package selected by sima-cli:

sdk-or-devkit$sima-cli neat install core

You can optionally install extras, including prebuilt tutorials, source-code tutorials, and tests.

Install a Selected Release

Use this option to install a specific Neat Library release. Append the version to the package name:

For Neat Library 0.1.0:

sdk-or-devkit$sima-cli neat install core@0.1.0

For Neat Library 0.2.0:

sdk-or-devkit$sima-cli neat install core@0.2.0

After Installation

On the DevKit, installation also provisions a PyNeat virtual environment at ~/pyneat:

devkit$source ~/pyneat/bin/activate

If you run the installer from inside an SDK container paired with DevKit Sync, the installer provisions PyNeat on the paired DevKit, not only inside the SDK container. PyNeat is installed into a virtual environment under ~/pyneat on the DevKit.

If you prefer building from source, follow the Build guide.

For SDK setup and DevKit Sync onboarding, see Neat Development Environment.

Install PyNeat in a Custom Virtual Environment

The Neat Library install artifacts include a PyNeat wheel. The standard installer uses that wheel to provision ~/pyneat for you. If you need a custom Python environment, install the downloaded wheel into a separate venv, conda environment, or other Python 3.11 setup.

Use this flow when the matching Neat Library runtime packages are already installed. The PyNeat-only target downloads the wheel, but it does not install or update the runtime .deb packages.

1. Download the PyNeat wheel:

   sdk-or-devkit$sima-cli neat install core -t pyneat

   To download the wheel for a specific Neat Library release, include the version:

   For Neat Library 0.1.0:

   sdk-or-devkit$sima-cli neat install core@0.1.0 -t pyneat

   For Neat Library 0.2.0:

   sdk-or-devkit$sima-cli neat install core@0.2.0 -t pyneat

2. Create and activate your target Python 3.11 environment:

   sdk-or-devkit$python3.11 -m venv ~/my-neat-envsdk-or-devkit$source ~/my-neat-env/bin/activate

3. Install the downloaded wheel:

   sdk-or-devkit$pip install ./pyneat-*.whl

The published wheel targets CPython 3.11 on linux_aarch64, so install it into a matching environment. For supported Neat Library, SDK, and DevKit software combinations, see the Compatibility Guide.

Neat CLI

The installed library provides the neat environment command. Run it from either the SDK or DevKit to see installed component versions, installed sima-cli playbooks, and whether newer artifacts are available.

In the SDK, the status output also includes Insight host-port mappings from $HOME/.insight-config/neat-port-map.json.

sdk-or-devkit$neat

Example output:

Neat Environment
  Mode               Neat SDK
  Sysroot            /opt/toolchain/aarch64/modalix
  Update check       online

Components
  Neat core              0.0.0+main-6735c35 channel=main latest=6735c35
  PyNeat                 0.0.0
  neat-runtime           0.0.1-main-e4b19553e07d
  neat-gst-plugins       0.0.1-main-e4b19553e07d
  neat-insight           0.0.0+main.74ae0b5 channel=main latest=74ae0b5 status=Running venv=/opt/neat-insight/venv
  Model Compiler        2.0.0.neat+main-1ebbc39 run activate-model-compiler to activate

Exposed Ports
  Insight Web UI     https://10.0.0.22:9900

  Name               Protocol Host Port (Start) Host Port (End)
  ------------------ -------- ----------------- ---------------
  mainUI             tcp      9900              -
  metadataUDP        udp      9100              9179
  rtsp.tcp           tcp      8554              -
  videoUDP           udp      9000              9079
  videoUI            tcp      8081              -
  webRTC             udp      40000             40199
  webSSH             tcp      8022              -

For automation and tool integrations, use JSON output:

sdk-or-devkit$neat --json

To update the Neat Library runtime, neat-insight, and installed sima-cli playbooks from the detected channel, run:

sdk-or-devkit$neat update

Next step

Continue with Hello Neat! to run your first Neat inference.