Skip to main content

Installation

As an AI application developer using Palette Neat, you have two setup options. You can develop directly on the DevKit, or you can work from a host-side Neat Development Environment and run on a paired DevKit. Both paths use the Neat Library and target the same DevKit runtime.

Prerequisites​

  • You must have a SiMa.ai Modalix 3.0 Developer Kit.
  • Install sima-cli using the installation guide in both your DevKit and host environments.
    • sima-cli is used to install the Neat Library and to set up the Neat Development Environment / DevKit workflow.
  • We recommend creating an account at community.sima.ai
    • This gives you access to assets such as precompiled models for testing. You can also ask and review questions in the forum.

For supported host operating systems, Neat Library versions, DevKit software versions, and Model Compiler compatibility, see Compatibility.

Start Installation​

Which installation should you choose?
For development simplicity

Develop on DevKit

Install the Neat Library and develop your app directly on a Modalix DevKit.

  • Fewer moving parts for initial setup.
  • Quick path to a first successful run.
  • Good fit for Python-heavy workflows and validating directly on the DevKit.
Install on the DevKit Direct-on-device setup
Recommended host workflow

Develop on host

Use the Neat Development Environment for faster builds, then test on a paired DevKit.

  • Builds in the Neat Development Environment are generally 5-10x faster.
  • Develop on the host, then run on a paired DevKit.
  • Better fit for larger C++ applications and agent assisted workflows.
Install Development Environment Host-side Neat setup

If your app is primarily Python, working directly on the DevKit is often sufficient because the Python runtime (pyneat) executes on the DevKit. Use the Neat Development Environment primarily when you need C++ cross-compilation, unified host/DevKit orchestration, or agent assisted workflows.