Can I use SEGGER’s development tools in my coding assistant? – Yes, of course!

The help of AI in development has already taken another step. We can not only use AI to tell it about our project idea and let it create the code. Development assistants can also assess their output, run tests, find bugs, and finally fix them. This works for PC applications and web services, and for embedded firmware, too.

Embedded firmware development and testing is more involved than for PC or web software. You need to connect to target hardware, flash the firmware image, control the CPU, and collect insights and outputs. With the SEGGER development toolbox (J-Link and J-Trace, Embedded Studio, Ozone and SystemView), we have already made traditional embedded software engineering simple and powerful. Now we can enable AI development assistants to benefit from these advantages, too.

The great news: You can already teach your AI to use the SEGGER development toolbox today. On our website and in our knowledge base you can find examples and getting-started guides for using the toolbox with readily available AI solutions.

Built for automation

In all our projects, a key objective has always been to enable and support reproducible automated workflows and external control. Every tool in the SEGGER development toolbox can be scripted and controlled from the outside:

• The J-Link SDK provides an API to all core functionality of J-Link, accessible in C/C++ applications and Python scripts.
• J-Link’s behavior can be adapted for any target device with J-Link Script Files.
• J-Link Commander can run automated sequences through Command Files.
• Embedded Studio provides command-line tools to automate building and testing.
• SystemView can be controlled from a terminal through command-line actions.
• Ozone provides a full scripting language to set up and interact with debug sessions from project files, the console, and with its Automation Socket Interface.

This is the foundation an AI agent can work with.

AI-assisted development in practice

We have put this to a test, talked to developers, and discussed with AI tool builders. We have seen AI assistants using the SEGGER development toolbox to autonomously find, analyze, and correct hard faults on embedded hardware, as well as AI agents which provided an in-depth analysis of the runtime behavior of firmware and made it understandable to non-engineers.

What’s next for AI-assisted development?

Our journey does not stop here.

Teaching an agent to do the right steps when a developer says “What is wrong with my firmware?” requires engineering, too. And while an intuitive GUI is useful for the human developer, an agent can be more efficient with structured text input and output from the tools.

We do not want every developer to reinvent the wheel and require a degree in prompt engineering. Our goal is to improve AI enablement, develop skills, and enhance our tools for use in AI-assisted development. Building on the toolbox we have today, the SEGGER AI Development Toolbox is going to provide valuable utilities and streamline development of professional AI solutions as well as bespoke developer setups.

Ready to get involved?

Check our dedicated AI page for news and updates, and follow this blog for the next posts in the series.