Designing a Keyboard PCB End-to-End with AI Using Claude Code and EasyEDA

I tested AI-assisted circuit and PCB design by having it produce a complete keyboard PCB, from component selection to layout, in about two hours. I share the full workflow, including the environment setup, the revisions I requested, and the areas where manual review was still needed.

■ Environment Setup and First Test
- Claude Code and EasyEDA's AI skill package from GitHub, working folder on OneDrive
- Simple battery-powered LED circuit as a trial run

■ Keyboard PCB Design
- Choosing a 60% keyboard with number keys and selecting the RP2040 microcontroller
- Tactile switches, diodes, and surrounding circuit components
- Layout revisions: QWERTY check, silkscreen labels, 180-degree rotation, footprint and outline adjustments
- Auto-routing results, component placement tuning, and manual adjustments

■ Manufacturing and Sponsor
- Ordering five assembled PCBs through EasyEDA and JLCPCB, totaling $66 with shipping
- JLCPCB flexible PCB options, with boards starting at $2 for 500x100mm

■ What Comes Next
- Keyboard firmware work carried out entirely through Claude Code

This is aimed at electronics makers and hobbyists who want to see a practical, start-to-finish example of AI-assisted design and where human intervention still applies. Watching gives you a concrete sense of the workflow, the tools involved, and what to prepare before trying it yourself. A follow-up showing the assembled keyboard in operation is planned, so results can be checked there as well.