Left-Behind Cabin Check
AI cabin inspection for rental fleets

Nothing gets left behind.

A small AI camera checks the inside of a car every time it's returned. It flags belongings left behind, litter and stains before the car goes out to the next customer.

Top-down camera view of a rear seat. Boxes mark a phone left on the seat, a stain on the middle seat, a cup and a wrapper on the floor. belonging 0.94 stain 0.81 litter 0.88 litter 0.72 CAM 01 · REAR CABIN · IR CAR RETURNED · CABIN EMPTY 4 FLAGGED · SENDING ALERT
Illustration of a post-trip check. Drawn for this page, not real model output.
The problem

Nobody checks the back seat until it's too late.

Rental cars change hands many times a week, and turnaround is fast. Staff check the outside for damage, but the inside gets a quick glance at best. The renter finds out their phone is missing at the airport, and the next customer finds the empty cup.

rental company

Return checks don't scale

Staff walk up to every returned car, often at peak hours. Items under seats and small stains get missed, and when a customer disputes a cleaning fee there's no photo to show.

renter

Lost items are hard to get back

By the time a renter notices a missing phone, wallet or passport, they may be on a flight home. Getting it back means calls, shipping and waiting.

next customer

The next customer inherits the mess

A wrapper or a spill missed at check-in becomes the next customer's first impression, then a complaint and a lower review.

How it works

A check after every trip, with nobody lifting a finger.

  1. Car is returned

    When a rental car is returned, or a trip ends, and the doors close, the system confirms the cabin is empty. It never runs with people inside.

  2. Camera captures the cabin

    A wide-angle camera with infrared lighting takes a picture of the seats and floor, day or night.

  3. AI inspects on the device

    A detection model finds anything that doesn't belong and labels it as a belonging, litter or a stain.

  4. The right person is told

    The rental desk, the customer or the fleet dashboard gets an alert with a cropped photo of what was found.

Even more accurate with a reference photo. When the camera is fixed in place, the system also compares each check against a photo of the clean cabin. Anything new stands out, which cuts false alarms from the car's own seatbelts, headrests and seat patterns.

What it finds

Three kinds of problems, each with its own alert.

belonging

Left-behind items

Things a customer will want back. They can be told before they've gone far.

  • phones
  • wallets
  • keys
  • bags
  • earbuds
  • glasses
  • kids' toys
litter

Trash

Things the next customer shouldn't find. Staff know which cars need tidying before they go out.

  • cups
  • bottles
  • cans
  • wrappers
  • tissues
  • receipts
stain

Stains and spills

Damage that needs cleaning, with a timestamped photo showing which rental or trip it happened on.

  • drink spills
  • food
  • mud
  • wet marks
Our approach

Built for the real inside of a real car.

Recognises the unexpected

People leave behind an endless variety of things. We use open-vocabulary AI models that can spot objects they were never specifically trained on, then sharpen them with real cabin photos.

Works in the dark

Infrared lighting keeps images consistent at night and in parking garages, where many rental cars are returned after hours.

Private by design

The check runs only when the cabin is empty, and the image is processed on the device. What leaves the vehicle is the alert and a crop of the item, not a video stream.

Moves to new vehicles easily

A new car model or a bus is mostly a new camera position and a new reference photo, not a new product.

Where it goes

Starting with rental cars, then anywhere people ride together.

First

Rental car companies

  • Return checks at branches and airport lots
  • Photo evidence for cleaning fee disputes
  • Faster lost-and-found for renters
  • Car-sharing and subscription fleets
Next

Public transport

  • Buses at the end of the line
  • Train and metro cars
  • Airport shuttles
  • Coaches and ferries
Then

Taxis and more

  • Taxi and rideshare drivers
  • Corporate and delivery fleets
  • Robotaxis, where no driver is there to look
Where we are

Early, and building in the open.

  1. Now

    Prototype

    Testing AI detection on photos of real car interiors to measure how well it finds items without any custom training.

  2. Next

    Real-world data

    Collecting cabin photos across many car models, seat colours and lighting conditions to train a model built for this job.

  3. Then

    Pilot

    Running the camera in a small number of cars with a rental car company.

Get in touch

Run a rental car company or a transit service?

We're looking for early pilot partners and people who deal with lost items and cabin cleaning every day. Tell us how it works for you today.