How it works

From goal to approved change.

This is the intended workflow. The proposal and policy steps exist as tested reference code today. The connected app and ECU steps are on the roadmap.

Set a goal Roadmap

For example: smoother cruise, cleaner cold starts, or a track day. You can describe it in plain language. The AI turns it into a typed intent, and you see exactly what it understood.

Log real driving Research · unverified

The app reads live data over the ECU vendor's documented, read-only interface. Which ECUs are feasible is still desk research. See the matrix.

Analyse with hygiene In development · simulated

Transients, warm-up and accel enrichment are filtered out. Lambda transport delay is compensated. Every sample carries a confidence. Cells without enough evidence are left alone.

Propose, then check In development · simulated

A planner turns the analysis into specific cell changes. The policy engine checks each one: change class, step limits, your role, and the safer direction. It rejects anything outside the envelope, however the request was phrased.

You approve In development · simulated

You see the diff, the rationale and the evidence. Nothing is applied without your explicit approval.

Apply Roadmap

On a third-party ECU, you apply the change in the vendor's own software. On our ECU, the change is re-validated on the ECU before it takes effect, and it can be rolled back.

What it won't do

Clear limits, on purpose.

  • It won't add ignition timing or boost on its own. Power-limit work needs a dyno and a professional.
  • It doesn't replace a professional tuner for a base map, a new build or anything near the limit.
  • It won't write to or flash a third-party ECU without an authorized, verified vendor interface.
  • It doesn't make a vehicle street legal or emissions compliant, and it won't remove emissions equipment.
  • It doesn't guarantee power, economy or safety outcomes. AI suggestions can be wrong. That's why you approve each one.