Commercial

Medkit Pro

Your whole fleet in one view. Faults come back with the fix.

The fleet layer on the open-source core: every robot and PLC behind one diagnostic API, and recurring faults matched to the fix that worked before.

Get access Start with open source

Apache 2.0 coreOn-prem & air-gapped

Medkit Pro faults dashboard - a three-vendor fault cascade across Allen-Bradley, Beckhoff and Siemens gateways, grouped by entity, with severity chips and live freeze-frame evidence — Faults dashboard · a three-vendor cascade on one screen

Self-healing loop

01Open core

Observe

Capture around the fault, not the firehose.

02Pro

Diagnose

Classify the root cause with fault correlation.

03Pro

Fix

Agent on rails: bounded, your model, human-approved.

04Pro

Verify

Health gates confirm the fix before the loop closes.

05Pro

Learn

Every fix becomes a reusable knowledge unit.

Fault evidence

Every fault carries its evidence.

One cross-vendor incident in one place: the PLC fault and the robot it stranded, side by side.

A freeze-frame of every relevant value, captured at the moment the fault broke.

A black-box recording of the seconds around the fault.

A fault document in Medkit Pro - the tank high-level fault with its course of events, and the freeze-frame values captured at the moment it broke — Fault document · course of events, state at the fault

AI assist

Answers from your plant, not the internet.

Press Explain and the panel carries the evidence with it - freeze-frame, live values, recording. The first time, the AI reasons and proposes. Your technician confirms what happened, and the answer becomes a record. When that fault returns, it answers in milliseconds, from your history, at zero AI cost.

The Explain panel in Medkit Pro - the AI names the three-vendor cascade as the root cause and proposes the fix, answered from a record the plant confirmed, at 70 percent confidence — Ask it · answered from a record your plant confirmed

What's included

Monitor the whole fleet

Every Medkit gateway - robots, PLCs, mixed cells - fans into one endpoint, and the whole site reads as one live faults view.

Triage with proven fixes

A deterministic matcher ranks causes against every incident before it - about 40% faster to root cause - and proposes the fix that worked last time. Your operator approves.

Wire in your AI

Chat over the live fleet with the assistant you already use, consent-gated, or a fully on-premise model. Nothing calls out on its own.

Ship updates that roll themselves back

Updates are signed, gated on live device health, and rolled back automatically if the gate fails. Stage campaigns across the fleet instead of driving to every site.

Go beyond ROS 2

Classic ECUs over UDS and DoIP, industrial protocols, and safety-certified runtimes - the same SOVD API serves them all. It already runs inside an ISO 26262 ASIL-B middleware on a safety RTOS.

Where it fits

One diagnostic layer for the whole stack, not one tool per vendor.

How to get it

Runs on top of the open core you already have. Start with a pilot on one mixed cell. License fleet-wide when it proves itself.

See selfpatch on your fleet

Tell us your stack - we'll show which of your recurring faults it could already resolve.

Get access Start with open source