Here are the current authoritative resources available regarding STM32Cube.AI (X-CUBE-AI):
1. Official Course Resources
1. ST Official Six-Lecture Course
· Content: Covers the entire AI development process, including data collection, model conversion (using STM32Cube.AI), deployment verification, etc.
· Features:
· Designed specifically for embedded engineers, integrating practical STM32 hardware experience;
· Includes live Q&A sessions (follow-up arrangements required).
· Access: Learn through official ST channels (such as the Electric Hall platform), noting that the course was released in 2019, and some tool versions may have been updated.
2. ST Edge AI Suite Practical Case Studies
· Resources: Provides over 50 edge AI case studies and 20+ documents (updated in October 2025), covering the entire process of model optimization and deployment.
· Tool Support: STM32Cube.AI serves as the core tool, enabling quick conversion of pre-trained models into STM32-optimized code.
· Access: Visit the ST official website (https://www.st.com) to obtain the toolkit and case library.
2. Recommended Books
1. “STM32Cube Efficient Development Tutorial” (2021)
· Highlights:
· Detailed explanation of the STM32Cube ecosystem, including integration methods for X-CUBE-AI;
· Provides development practices from basic to advanced, suitable for systematic learning.
· Accompanying resources: Source code, videos, etc., can be downloaded from the publisher’s platform.
2. “Fundamentals and Project Practice of STM32 Microcontroller – Micro Course Edition” (2nd Edition, 2025)
· Target Audience: University students and beginner developers;
· Content: Incorporates fundamentals of AI development, with practical training projects and ideological elements;
· Resources: Micro-course videos + teaching materials (Tsinghua University Press).
3. Extended Learning Channels
· ST Official WeChat Account:
Regularly publishes tool updates, case studies, and interactive activities (such as development board giveaways), allowing timely access to resources.
· Corporate Training Courses:
· Institutions like Huqing Vision and Maker Academy offer specialized embedded AI courses (including practical applications of STM32Cube.AI);
· Covering model deployment, optimization, and IoT application development (course resources need to be verified for timeliness).
Currently, there are no official publications on STM32Cube.AI by ST, but the above resources cover the complete path from tool usage to project implementation. If further course selection or debugging guidance is needed, specific application scenarios can be provided.
Step 1: Installation
“SetupSTM32CubeMX-6.15.0-Win.exe”
Step 2: Install AI Support Package
“STMicroelectronics.X-CUBE-AI.10.2.0.pack”
Step 3: Install
“st-stm32cubeide_1.19.0_25607_20250703_0907_x86_64.exe”
Step 4: Install Python 3.9 (3.10)
Step 5: Install TensorFlow, run
cmd – pip install -i tensorflow https://pypi.tuna.tsinghua.edu.cn/simple
STM32CubeMX and AI Deployment
The core purpose of this series of installation steps is to deploy AI models (edge AI/TinyML applications) on STM32 microcontrollers, analyzed as follows:
1. Basic Hardware Configuration
STM32CubeMX (Step 1) is the official MCU configuration tool from ST, used to generate hardware initialization code (clock/pin/peripheral configuration), providing underlying hardware support for AI model deployment.
STM32CubeIDE (Step 3) is the integrated development environment that supports code compilation and debugging for STM32.
2. AI Model Conversion and Optimization
X-CUBE-AI (Step 2) is the core AI extension package, with functions including:
✓ Converting models trained with frameworks like TensorFlow/Keras into C code executable on STM32
✓ Automatically optimizing model structures (quantization/pruning) to fit the limited computing power of MCUs (such as the Cortex-M series)
✓ Providing AI inference runtime libraries (supporting common networks like CNN/RNN)
3. Software Ecosystem Dependencies
Python 3.9/3.10 (Step 4) is the runtime environment for TensorFlow, used for:
✓ Model training and validation (to be completed in advance)
✓ Generating model files (such as .h5 or .tflite format) for X-CUBE-AI conversion
TensorFlow installation (Step 5) further validates that this is an end-to-end AI workflow: from PC-side training to MCU-side deployment.
4. Typical Application Scenarios
Edge AI Inference: Running lightweight AI models (such as sensor data classification, voice wake word detection) on resource-constrained STM32
TinyML Development: Complying with TinyML technical characteristics (low power consumption, small memory footprint, real-time requirements)
Industrial Cases: Motor anomaly detection, smart home voice control, wearable device biosignal processing, etc.
Note: For deploying more complex models (such as object detection), it may be necessary to combine with the STM32MP series (MPU with Linux) or external AI acceleration chips.