Kevin.
Docs
Getting startedReviewing photo sets

Reviewing photo sets

Kevin proposes the grouping; you approve it before anything is identified.

In this article

Between upload and processing, Kevin proposes photo sets: shots taken within seconds of each other, in the same place, are almost always the same object. Three angles of a sofa become one line item rather than three. A lone photo is simply a set of one.

This screen runs before identification. It shows filenames, timestamps, and how far apart the shots were — not item names, because nothing has been identified yet.

What you can change

  • Merge — select two or more sets that belong to the same object and combine them.
  • Split — break a set apart when Kevin grouped two different objects shot back to back.
  • Skip — mark context shots, blurry frames, or anything you do not want counted.

The tally

The counter shows photo sets, not line items — 50 photos might propose 47 sets. The item count is not known until identification runs, and it will be lower still once unmatched and context photos fall out.

Notes that steer identification

You can attach a note of up to 120 characters to any set. It rides on the whole set and steers what Kevin looks for: "the TV is the Sony, not the Samsung," or "this is the mahogany one, not the veneer." Notes influence identification only — they never affect price.

Notes are pre-processing only. If an item comes back misidentified, correct it by editing the worksheet row, not by re-running identification.
Still stuck? Email kevin@kevin.co — it goes straight to a person who has settled claims.