JumperFree

Wire less.

Let AI help you build real circuits
the way it already helps you write and debug software.

After code.

The firmware is ready.
The real circuit still asks you
to hunt pins, place jumpers,
and check the same paths again.

An agent reads the ask. A deterministic router plans the nets. An STM32 controller carries them out. Each layer stays in its lane—so the build stays explainable and checkable, not a model guessing at switch registers.

Loop.

“Route the sensor’s SDA and SCL
to the board’s I²C pins.”

Netlist
  SDA: [clip_01, clip_09]
  SCL: [clip_02, clip_10]

Speak

Describe the pins you want linked in plain language. The agent turns that intent into a structured netlist the software can check.

Watch

Software validates the nets and computes a route. RGB LEDs mark clips and nets so you place parts by the light, not by guesswork.

Connected

The verified plan rides USB to the controller, which closes the physical paths. Run, read control state, then shape the next link.

Capable.

Route

From a netlist, a deterministic router computes the connection plan. Every link has a path you can follow back.

Switch

STM32F103 and CH446Q drive the crosspoint. Break-before-make keeps the plan aligned with the real electrical path.

Light

RGB LEDs track clips and nets, mapping the software graph onto the holes under your hands.

Host

USB to a laptop. Python CLI or API to orchestrate circuits inside the tools you already use.

Agent

Natural language becomes a structured netlist, then deterministic software takes over. The agent never writes switch registers.

Safe

Isolation, default-open paths, and an emergency all-off draw a clear safety boundary. Control feedback reports switch state; meters still confirm live electrical values.

Join.

If this is the build loop you want,
leave an email and stay close to JumperFree.

We use your email only for JumperFree updates.