What the camera sees, the system can't unsee.
Check-in / check-out photos run through an OCR model trained on dash-cluster imagery. The digit string is hashed to the VIN at upload — readable forever, editable never.
Each layer is independent. Each layer is auditable on its own. Together they leave no room for a he-said-she-said.
Check-in / check-out photos run through an OCR model trained on dash-cluster imagery. The digit string is hashed to the VIN at upload — readable forever, editable never.
Hosts enter odometer, fuel/SOC, and damage state on the same form their phone uses for the photo. Entries are timestamped server-side and reconciled against the OCR read.
Smartcar / OBD streams odometer, speed, location every ~60s. Continuous data resolves any photo-vs-host disagreement automatically — and feeds the anomaly engine in real time.
A simulated Phoenix-metro trip plays below. At mile 21, the telemetry layer flags a speed reading no Camry can produce.
The packet attaches itself to the claim — no host follow-up, no adjuster waiting. The anomaly flag is the first thing on the page.
Hosts get protection. Guests get speed. Insurers get certainty. Nobody else in P2P does this.
The forensic stack isn't just better evidence — it's longer-retained, harder-to-alter evidence than the law requires.
Arizona's P2P car-sharing statute sets minimum record-keeping, insurance, and disclosure rules. ItWhip's forensic stack exceeds them on every axis — particularly retention, immutability, and proof-of-mileage.
Every trip's photo + host entry + telemetry stream is encrypted and held for 6 years from trip end. Adjusters can request export at any point.
Each trip event is signed and chained to the prior block. A single byte tampered downstream invalidates the whole tail — visible to any auditor with read access.