Skip to the page content
Parcel OCR
Open the officer app

  1. A parcel arrives. Nobody types it in.

    At high-volume properties, the front desk registers 1,000+ parcels a day by hand, about 45 seconds each (reported in the requirements). Parcel OCR does the typing.

    Open the officer appSee how it works
    Line drawing of a condominium tower cut open at one corner. In the lobby at its base, a courier hands a parcel across the counter to an officer.

  2. The camera reads the label.

    Capture with the webcam or upload a photo. The label is read inside your organisation's own systems, and every field gets a box and a confidence score.

    Close-up of the counter: a webcam on a stand shines a lavender beam onto the label of a parcel, and blue boxes outline the fields it reads.

  3. One unit. One recipient. Or none.

    The label is matched only against residents of this property. A similar name never matches on its own, and if two units fit, neither is chosen.

    The same tower drawing with a blue line from the lobby counter up to one unit door; the rest of the tower is faded.

  4. When it is unsure, it waits.

    Anything uncertain goes to the hold rack with the reason in plain words. Nothing is guessed: an officer checks it, fixes what is marked, and confirms.

    The lobby with its wall of cubbies. One cubby glows amber under a question mark while the officer speaks with the courier.

  5. Saved once. Notification queued.

    Confirmed by an officer, or automatically when the property allows it, the parcel is saved with one pending notification. Retries never create a second parcel.

    The lobby again. One cubby glows green under a check mark, and a bell icon sits beside the officer at her desk.

Walk a parcel through the officer app.

Capture, matching, the hold rack and saving work today in this preview. Sending notifications and the real recognition model come later.

Open the officer appSee how it works

What is in this build

  • Capture by webcam, image upload or auto-capture
  • The label image with a box and a confidence score for each field
  • Unit and recipient matching inside one property
  • Duplicate tracking numbers held before saving
  • Hold rack and queue, with an owner for each job
  • One parcel, one record, with a notification queued

What comes next

  • Sending notifications to residents
  • The real text recognition model
  • Real resident data and sign-in
  • Dashboards and property reports
  • Collection and handover flows