The problem
Most vehicle issues announce themselves first as a sound. Drivers know something is off, but they have no good way to describe it to a mechanic — and no way to know whether it’s a 30-minute fix or a 3,000-dollar one before the shop quotes them.
The product
MyAutoWhiz is a vehicle intelligence platform built around an audio-diagnostic mode. You record the sound — engine bay, wheel, transmission, brakes, idle — and a classifier returns the most likely categories of cause, with a follow-up checklist that tells you what to confirm before paying for a diagnosis.
How it works
- Record. Guided capture for the most useful diagnostic angles (cold start, idle, slow roll, turning lock-to-lock).
- Classify. An ML classifier returns a ranked list of likely causes from common automotive failure categories.
- Triage. Each cause links to a short checklist — visible cues, a simple test the driver can run, an honest range of repair cost, and the right shop type to take it to.
- History. Recordings are kept per vehicle so you can compare a current sound against the same spot a month ago.
Stack
- iOS: SwiftUI with native audio capture and waveform visualization.
- Audio ML: model inference via a backend service, with on-device preprocessing for noise reduction and segmentation.
- Backend: lightweight Python service for inference, recording storage, and per-vehicle history.
- Distribution: App Store with StoreKit 2 for the paid tier.
Status
MyAutoWhiz is in development.