Via — AI Robotic PCB Probing

Hantek measurement hardware beside an automated robotic PCB probe station with a fine needle over a populated circuit board.

“We do not assume the circuit—we verify it.”
— in voce Dr. a.Eisler

Purpose

Via is a low-cost, AI-operated robotic probing system for physical interrogation of PCBs, automated measurement workflows, and AI-guided diagnosis of electronic faults.

System

Motion (precision probe positioning), sensing (electrical measurement), and reasoning (diagnostic planning via the Eisler engine) are combined behind a single bounded Via API—so automation touches the board through safe, reviewable actions.

Motion — G-code gantry (e.g. repurposed 3D printer)
Measurement — DMM / oscilloscope drivers
Board model — nets, pad coordinates, and alignment with the KiCad layout (via kicad-mcp, below)
Eisler — planning and reports; no direct motor control

Layout

We run kicad-mcp as a Model Context Protocol bridge to KiCad. The AI can query the open board the same way a human would browse the layout: nets, footprints, pad geometry, copper layers, and design rules—grounded in the source .kicad_pcb instead of a one-off export. That lets Eisler plan where to probe and what should be electrically consistent before the gantry moves.

Repository

Design, Python API, and hardware notes live in the open-source repository.

View on GitHub →