This project is a physical four-track audio recorder that explores how hardware and software can be developed together through AI-assisted workflows. The system consists of a simple Arduino-based controller connected to a macOS application, enabling audio recording and track layering through physical buttons and LED feedback.
A tap on the TRACK button moves between tracks while a held press arms one for recording. The MAIN button captures audio or erases it. Yellow LEDs breathe gently when armed, blink urgently when recording, and glow steady when holding audio. The red LED stands watch—illuminating only when recording is live.
The defining aspect of the project is that all the software was created by Claude. Claude wrote the complete Arduino firmware, designed the macOS application, and handled the logic that connected physical inputs to digital recording behavior. My role focused on wiring the hardware and testing interactions, while Claude iteratively generated, debugged, and refined the code across both platforms.
The idea emerged from an interest in understanding how quickly functional hardware–software systems could be prototyped when assisted by AI. Instead of separating hardware and software development, both were built in parallel, allowing for rapid experimentation and immediate feedback.
The project was completed within half a day, demonstrating how AI can dramatically compress development timelines. This process shifted the focus from writing boilerplate code to understanding system behavior, testing edge cases, and refining interactions.
