Caption photos at the speed of recognition.

idXpresso identifies people in your photos with private, on-device facial recognition and automatically writes their names and titles straight into your image metadata — built to drop seamlessly into a professional editorial and DAM workflow.

100% on-device Free to try macOS 13+ · Apple silicon
Private, on-device recognition Writes standard JPEG metadata Built for editorial & DAM Learns as you work

Everything you need to caption faster

Purpose-built for photographers and metadata specialists who name people at scale.

Privacy by design

All recognition runs locally. Face encodings and your identity database stay on your Mac — never uploaded to any server.

Automatic metadata

Names, titles, and keywords are written or prepended directly into industry-standard JPEG metadata fields as you go.

Identity management

Build and maintain your own rapidly searchable database of people — the more it grows, the faster you caption.

Workflow integration

Keyboard-centric navigation and low-friction interaction designed to sit naturally alongside your existing photo apps.

Learns over time

Recognition improves as you train it and as your identity library grows, getting sharper with every session.

Built for deadlines

Optimized for offline work, deadline production, and batch encoding — reduce the most time-consuming part of captioning.

From import to captioned in four steps

Keep full editorial control while idXpresso handles the repetitive work.

1

Import photos

Point idXpresso at a folder or watch folder and it scans for faces on-device.

2

Recognize people

Matches faces against your private identity database and surfaces likely names.

3

Confirm & refine

Approve, correct, or add identities with fast keyboard navigation. It learns from every choice.

4

Write metadata

Names, titles, and keywords are written into the photo's standard metadata, ready for delivery.

Let the photo write its own caption

This isn't only about identifying people you don't know. You bring the what, where & when; idXpresso recognizes the faces and completes the caption with the who — then writes it everywhere it belongs.

You prepare
The what, where & when
A general caption added on ingest — the scene, the place, the date.
idXpresso completes
The who
Recognition names the faces it knows and finishes the caption for you.

Even if you know everyone — even with a macro or two up your sleeve — watch how the time to caption a group of 5 compares:

Type it by hand
At ~40 WPM — the average for an everyday adult typist¹ — entering the full name & title across all three fields.
Text macro / snippet
Fast — but only for people you've pre-built a snippet for, and you still have to recognize each face yourself.
idXpresso auto-match
For a face it auto-matches, you just confirm and commit — and it names the person for you. Faces it misses fall back to the quick picker.
And one commit writes the name into all three standard fields: CaptionKeywordsPersons shown

Estimates for illustration, and only for faces idXpresso auto-matches: name & title such as “John Doe, CEO of Any Company Global Inc.” (~40 characters) entered across the caption, keywords, and persons-shown fields. Typing based on a 40 WPM average; your mileage will vary. Recognition won't catch every face — misses drop to the quick picker — but macros can't identify a face at all, and idXpresso can.

Start free. Upgrade when you scale.

Try idXpresso with no commitment, then unlock the full professional toolkit with Pro.

Free

idXpresso

Free · up to 25 identities
  • Private on-device recognition
  • Build an identity database
  • Keyboard-centric captioning
Pro · 7-day free trial

idXpresso Pro

$9.99/mo or $99.99/yr
  • Unlimited identities
  • Metadata writing & batch encoding
  • CSV import/export, watch folder, learning mode

Pro is an auto-renewing subscription billed through your Apple ID. See in-app pricing for current rates in your region. Cancel anytime in your Apple ID settings.

Stop typing names. Start shipping captions.

Download idXpresso for Mac and let recognition do the repetitive work.

Requires macOS 13.0 or later · Apple M1 chip or later