How to install X-CUBE-AI through STM32CubeMX

This article describes how to install the X-CUBE-AI Expansion Package through STM32CubeMX.

Info white.png Information
  • X-CUBE-AI is a software that generates optimized C code for STM32 microcontrollers and neural network inference. It is delivered under the Mix Ultimate Liberty+OSS+3rd-party V1 software license agreement (SLA0048).

X-CUBE-AI is an STM32Cube Expansion Package that expands the capabilities of STM32CubeMX and is a part of the STM32Cube.AI ecosystem. It offers the automatic conversion of pretrained artificial intelligence algorithms, which include neural network and classical machine learning models. Additionally, it integrates a generated optimized library into the user's project.

1. Prerequisites

  1. STM32CubeMX (v6.9.1 or higher) installed on your computer
  2. myST account
  3. Internet access to download the X-CUBE-AI package
  4. Proxy configuration (optional)

2. Proxy configuration (optional)

If the PC running STM32CubeMX is connected to a computer network using a proxy server, STM32CubeMX must connect to that server. When the connection is established, STM32CubeMX can access the internet, get self-updates, and download firmware packages. Proxy configuration is typically (but not always) required on enterprise networks.

1. Launch STM32CubeMX and click [Help] > [Updater Settings]

CubeMX Proxy

2. Under the Connection Parameters tab, choose the appropriate Proxy Server Type

Fill in the proxy information, if needed. To ensure that STM32CubeMX can access the internet through the configured proxy, press [Check Connection]. Once the connection has been verified, press [OK].

If needed, contact your IT administrator for proxy information (proxy type, http address, port). For detailed information about connection parameters, see the section 3.4.2 in STM32CubeMX user manual[1].

CubeMX Proxy

3. X-CUBE-AI installation

Once the STM32CubeMX installation is complete and STM32CubeMX can access the internet, follow these steps to install the X-CUBE-AI package:

1. Launch STM32CubeMX and press [myST] -> [Login]

CubeMX Launch

2. Enter your myST account email address and password, and press [Login]

If you do not have a myST account, press [Create Account] to make one'.

CubeMX Login

3. Once you are logged in successfully, press [Help] -> [Manage embedded software packages]

CubeMX Manage

4. Open the Embedded Software Packages Manager window

Press the [Refresh] button to get an updated list of add-on packs. Go to the [STMicroelectronics] Tab, press [X-CUBE-AI], select the [Artificial Intelligence] package version you wish to install, and click the [Install] button.

CubeMX Select

5. Follow the installation steps on the screen and click [Finish]

CubeMX Accept

6. Once the installation is successful, the X-CUBE-AI Artificial Intelligence package is marked with a green box

Click [Close] before proceeding.


7. In STM32CubeMX, click [File] > [New Project] and select your preferred MCU from the [MCU/MPU Selector] tab, or select your preferred board from the [Board Selector] tab

Then, under the [Pinout & Configuration] tab, press [Middleware and Software packages], and then select [X-CUBE-AI].


8. In the Software Packs Component Selector window, select STMicroelectronics.X-CUBE-AI

Then enable Core: Artificial Intelligence X-CUBE-AI -> Core -> Enable Finally, select the application type: Device Application -> Application -> Select "SystemPerformance". Click on [Ok] when done in the Software Packs Component Selector window.

Note that X-CUBE-AI currently supports three application types, namely SystemPerformance, Validation, and ApplicationTemplate. See the sections 9, 10, and 11 in X-CUBE-AI getting started manual[2] for more details.

X-CUBE-AI Enable

Since this is the first use of the X-CUBE-AI Expansion Package, an OS-specific part of the package is downloaded and installed automatically.

X-CUBE-AI Download

9. Once the download is complete, X-CUBE-AI is ready for use in your project


For more information on how to use X-CUBE-AI in your project, refer to the section 3 in X-CUBE-AI getting started manual[2].

4. References