KiCadAI
Go CLI toolkit for AI-assisted KiCad design that generates, inspects and validates projects, with an ERC/DRC feedback loop and an agent skill contract.
Facts
- Builder
- dshills
- License
- MIT
- Connects to
- KiCad
- Repository
- https://github.com/dshills/KiCadAI
Repository stats
What it does and how it connects
KiCadAI is a Go CLI toolkit that turns bounded behavioral requirements into deterministic, evidence backed native KiCad projects, and refuses when it cannot prove a design inside its reviewed component, model, simulation, and physical design envelope. Its stated rule is that AI proposes and KiCadAI proves or refuses.
The input names behavior only: the schema rejects component, topology, model, internal net, value, geometry, route, provider, and repair instructions. Deterministic code then owns architecture and part selection, calculations, simulation, schematic and PCB generation, placement, routing, validation, and the final pass or refuse decision, with KiCad ERC, strict DRC, and normalized round trip checks on the written .kicad_sch, .kicad_pcb, and .kicad_pro files.
The featured demo in the README is a protected 0.1 A/V programmable current output, recorded on KiCad 10.0.3, where the tool derived candidate architectures, ran simulations and corner evaluations, selected catalog parts, and generated the same native project twice with identical hashes. The README states this is an ERC/DRC checked deterministic candidate, not a fabrication release, and that human electrical, thermal, mechanical, compliance, and manufacturing review remains required.
Alternatives
Related Error Index entries
Sources
3 SourcesNot verified
- No independent user voice found: recent repo activity is maintainer pull requests.
- Pricing: none stated on the repo.
Last verified: 2026-11-02. Listings are informational and are not endorsements, rankings, or compatibility claims.