After multiple failed attempts, I finally got Claude Code to successfully modify an STM32 project - enabling touchscreen functionality and updating the display on touch events. The difference? Better context setup and prompting strategy.
In this video, I walk through:
1. What went wrong in my previous attempts
2. How I improved my prompting approach
3. Setting up proper context for Claude Code
4. The actual STM32 touchscreen implementation
5. Key lessons learned for working with AI coding tools on embedded projects
If you've struggled to get AI coding assistants to work with embedded systems or hardware projects, this video shows what actually made the difference.
Tools used:
1. Claude Code
2. STM32 microcontroller
3. CodeForge vscode extension https://marketplace.visualstudio.com/items?itemName=TulipTreeTechnology.codeforge
Chapters:
00:00 Introduction to AI-assisted embedded development workflow
00:27 Setting up project structure and slash commands
02:06 Creating a prime command for project context
04:17 Generating automated README from codebase
07:15 How slash commands improve AI context management
08:13 Initializing Code Forge development environment
10:17 Setting up MCP servers for Docker integration
13:13 Building firmware and fixing ARM toolchain issues
16:41 Troubleshooting ST-Link board flashing problems
21:06 Converting images to C arrays for display
24:34 Debugging blank screen and frame buffer optimization
28:51 Key lesson: Context preparation makes AI succeed
#ClaudeCode #STM32 #EmbeddedSystems #AIcoding #Microcontrollers #CodeForge
