It learns which frames you star.
A new shoot then arrives already picked. On your Mac, from your own catalog. Your photographs never leave your disk.
How it works
Point it at your photos
Sieve finds your Adobe Photoshop Lightroom Classic catalog on its own, or you hand it a folder of originals. It reads names, capture dates, and the stars and rejects you have already given. It writes nothing back until you ask it to.
It learns from your stars
Every frame you have rated is a label. Sieve renders your originals the way the camera profile intends, measures each one with a vision model that runs on your Mac, and fits a small model on your marks. Press Learn again whenever you have rated more.
New shoots arrive picked
Open a set you have never looked at and the frames Sieve believes you would star are already under the light, bursts collapsed, the rest in capture order below. With the plug-in installed, the picks show up in Lightroom Classic as a collection.
What we can honestly say
On its maker's own library of about 8,000 unedited raw frames, Sieve put about half of the frames he later rated 4 or 5 stars into a top-10% pick, twice what chance would.
That is one library. Yours will differ, and the app shows you its own numbers before you trust it. Nobody else's library has been measured yet, so there is no second number to give you. When there is, it goes here.
Before you sign up
- Price
- Planned as a one-time purchase with a year of updates. The number is not set yet, and the waitlist is free.
- Mac only
- macOS 14 or later, Apple silicon. No Windows and no Intel build.
- Editors
- Lightroom Classic today, through a plug-in. Your marks also reach Capture One, Bridge, darktable and Photo Mechanic as XMP sidecars.
- What it needs from you
- Photos you have already rated. Below about 300 rated frames there is nothing personal to learn, and Sieve says so instead of guessing.
- Off the network
- No account and no upload. The app makes two kinds of connection: the one-time model download and an update check.
- What it is not
- Not a culler. It does not read faces, closed eyes or focus. It ranks by the taste your own stars describe.