Dear all,
The wiki page for STM32Cube AI Studio (at https://wiki.st.com/stm32mcu/wiki/AI:STM32Cube_AI_Studio_documentation) states that the software is compatible with “STM32CubeIDE, IAR, Keil, GNU Arm Embedded Toolchain”. However, when creating a new project it seems that at least STM32CubeIDE is required, and there does not appear to be straightforward support for using the bare GNU Arm Embedded Toolchain (for example, when integrating with STM32Cube for VS Code via CMake).
From a developer perspective, especially for those of us working with CMake‑based workflows and VS Code, proper and maintained support for the standalone GNU toolchain would be extremely valuable. It would allow tighter integration with VCS and existing CI/CD pipelines, easier cross‑platform setups, and would avoid forcing a switch to a specific IDE just to generate or maintain AI‑enabled projects.
Could the STM32Cube AI Studio team clarify the current level of support for the GNU Arm Embedded Toolchain outside STM32CubeIDE, and whether there are plans to:
Officially support and document a CMake‑friendly flow targeting the bare GNU toolchain (e.g. for STM32Cube for VS Code).
Keep the GNU toolchain integration updated in step with STM32CubeIDE and the Neural‑ART / STM32N6 toolchain evolution.
Provide example projects or templates showing how to consume the generated AI libraries in a pure GNU+CMake+VS Code environment.
This kind of documented, first‑class GNU toolchain support would greatly help adoption in more flexible, modern development setups and make it easier to bring STM32‑based Edge AI into existing codebases and workflows.
In the meantime, could the STM32Cube AI Studio team, or fellow developers facing the same issue provide instructions for possible workarounds?
Thank you in advance for your kind support.
Best,
Walter
