idXpresso mascot presenting
Part 1

Getting to know the interface

A tour of the main window — every panel and control, plus the special views you'll use to identify people and write captions.

Section 1 · The main window

Main window's initial view

idXpresso is organized into four stacked areas, top to bottom. Hover or tap the boxes and controls for details, and click a section name to jump to its explanation.

Open recognition mode (the main window) anytime with R. And if all windows are closed, opening a photo into idXpresso automatically launches this window once the photo is processed.
Hover or tap the boxes and controls for details · click a section name to jump.
1/1
Face 1Ignored
Face 2
John Doe (90%)
COO John Doe
Face 4
Face 5
Commit
The idXpresso main window, top to bottom.
Section 2 · Special views

Main window's special views

Beyond the default layout, the main window has three focused modes. Switch between them below and hover the highlighted controls to see what they do.

Hover or tap the highlighted controls for details.
1/1
Caption Write Metadata
Director Jane Doe and COO John Doe Jane walk with others from a helicopter.
Commit
Caption editor view.
1/1
Face 1Ignored
Face 2
John Doe (90%)
COO John Doe
Face 4
Commit
Watch Folder Idle
Enabled
Configure…
Watch folder controls — opened from the folder icon.
1/1
Face 1
Face 2
John Doe (90%)
COO John Doe
Face 4
Commit
Suggested Matches
⌘1Jane Doe88%
Other Known People
John Doe14%
ID picker — opened from a face's name dropdown.
Add Identity
Jane Doe
CEO of Some Company Jane Doe
Photo URL or path (optional)
Bio URL (optional)
Cancel Create
Add Identity — opens when you add a new name from the picker.

Caption editor

Switches the lower pane from the face list to a full caption editor for the current photo, so the people you named flow straight into a finished caption.

Caption field

Type or edit the photo's caption. Names you confirmed can be pulled in as you write.

Write Metadata toggle

When on, the caption is written into the file's standard metadata when you commit. What gets written is configured in Settings ▸ Metadata.

View toggle

The document icon in the status bar switches between the face list and this caption editor I. Its color and fill tell you the caption's state at a glance:

Orange — read-only metadata. The app can't write it; on commit you're offered to copy names / captions to the clipboard.
Green — has pre-written caption info.
Blue — metadata completed through facial recognition.
Solid = caption mode  ·  Outline = face list mode.

Watch folder controls

Point idXpresso at a folder and it processes new photos automatically as they land. Open the controls from the folder icon in the status bar.

Status

Shows whether the watch folder is idle or actively processing incoming photos. The folder icon in the status bar changes color to reflect this:

Green — watch folder is on: set, active, and monitoring.
Orange — error.

Enabled

Turn automatic watching on or off without losing your configuration.

Configure…

Choose which folder to watch and how newly added files are handled — the full options live in Settings ▸ Watch Folder.

ID picker & adder

Click a face's name dropdown to open the picker. Search your database, apply a suggested match, or add a brand-new person on the spot.

Search or add

Find an existing identity by name, or type a new name to create one.

Suggested matches

Ranked candidates with confidence — press 1 to apply the top pick.

Other known people

Everyone else in your database, each with its own match confidence.

Add a new identity

Type a name that isn't in your database to open the Add Identity dialog (switch to Add identity above). Set the name and a title/caption, add optional photo and bio links, then Create to save it and apply it to the face.

Section 3 · Identity Manager

The Identity Manager

The ID Manager is your people database. Browse and search identities, add or merge them, attach photos to build face encodings, and see each person's encoding status at a glance. Hover or tap the controls to explore.

Open the Identity Manager anytime with I.
Search identities
A. Person
Another Person
J. Doe
CEO of Some Company Jane Doe
Jane Doe
Director
John Doe
COO
J. Doe Edit Delete
CEO of Some Company Jane Doe
test.com Error: Invalid URL Process
CreatedJul 2, 2026
UpdatedJul 2, 2026
Drop photos here to add encodings
5 2 1
Add Encodings for J. Doe 0 of 4 selected
multi_face.jpg Face 1/4Conf: 88%  Sharp: 534
multi_face.jpg Face 2/4Conf: 88%  Sharp: 356
multi_face.jpg Face 3/4Conf: 85%  Sharp: 306
multi_face.jpg Face 4/4Conf: 80%  Sharp: 91
Cancel Add 0 Encodings

When a dropped photo contains more than one face, idXpresso asks which face(s) belong to this identity. Each candidate shows its detected-face number, match confidence, and image sharpness — pick the right one(s) to add as encodings.

The Identity Manager — your searchable people database.

Understanding encoding status

Each identity shows an icon for its encoding state — whether idXpresso has usable face data for that person. The same states are tallied in the bottom-right status bar.

Grey — linked, not processed

The identity has a linked photo (for example a URL) that hasn't been processed yet. Process it to build an encoding.

Green — encoded, not verified

Has an encoding that's used for matching, but you haven't verified it yet — for example an encoding built from a link in an imported CSV.

Blue — encoded & verified

Has an encoding used for matching that you've also verified. The most trusted state.

Red — error

Something needs attention. The specific error is shown in the identity's detail panel — fix it and re-process.

Section 4 · Learning Mode

Learning Mode

Learning Mode is a fast way to train idXpresso on people. Feed it captioned photos and it both creates the ID records and builds their face encodings in one pass. Hover or tap the controls to explore.

Open Learning Mode with L. For clean encodings, each learning photo should contain a single face (or otherwise be high quality).
Learning Mode
Add Cancel Clear Learn
STATUSFILENAMENAME WITH TITLE
Processing file_1.jpg
Encoding file_2.jpg John Doe COO John Doe
Valid file_3.jpg Jane Doe Director Jane Doe
Processing 2 of 3 3 photos 2 matched 2 learned
Learning Mode mid-scan — each row moves Processing → Encoding → Valid. Watch the status line, bottom-left.

Drop photos

Drag captioned image files straight into the window.

Use Add

Click the Add button to open and pick photos.

Caption trigger

A metadata caption can start Learning Mode automatically (see below).

Caption format

Learning Mode reads each photo's IPTC caption to know who's in it. Use this pattern:

NAME;NAME & TITLE

Example: Jane Doe; Director Jane Doe → creates an identity with Name = Jane Doe and Name with title = Director Jane Doe.

  • Name only. A caption with just Jane Doe creates the entry with only the name (no title).
  • Force Learning Mode. Prefix the caption with idxpresso: — e.g. idxpresso: Jane Doe; Director Jane Doe — and idXpresso switches to Learning Mode automatically, whatever mode it's in. You can drop a learning photo into Recognition mode and the app knows to switch.
Section 5 · CSV Import

Import identities from a CSV

Bring in a whole list of people at once. Map your spreadsheet's columns to idXpresso's fields, preview the result, and import — creating ID records in bulk. Hover or tap the controls to explore.

Map CSV Columns
Map CSV Columns · 2 rows
Name *
Name
John Doe
Name with Title
Title
COO John Doe
Bio Link
Not mapped
Photo Link
Not mapped
NameTitle
John DoeCOO John Doe
Jane DoeDirector Jane Doe
First Row is Header Encode URL(s) Cancel Import
The CSV importer — map columns, preview, then import.
idXpresso mascot presenting
Part 2

Understanding how idXpresso works

What the colors, statuses, and confidence scores mean — how idXpresso decides who's who behind the scenes.

Section 1 · Bounding boxes & match classes

Understanding bounding boxes & match classes

Every detected face carries a color that signals its match status. The same color appears on the bounding box in the photo and on the accent bar in the face list, so a face reads the same way in both places.

Face 2 ⌄

Detected Yellow

A face was found but nothing cleared the auto-match confidence threshold set in Settings ▸ Recognition, so it's left for your review. This class still includes suggestions — sometimes high-confidence ones that just missed the threshold. Open the ID picker to choose: when idXpresso is confident enough it shows a confidence-ranked list, otherwise it falls back to an alphabetical / default sort (set by Sort identity suggestions by).

John Doe (90%)Auto-matched

Auto-matched Green

idXpresso proposed a confident match from your database, shown with its confidence. Confirm it with thumbs-up, or correct it if it's wrong.

John DoeConfirmed

Confirmed Blue

You've verified and applied this identity — the face's metadata is complete through facial recognition and ready to commit.

Face 1Ignored

Ignored Grey

You marked this face as ignored. It's skipped and won't be written into the caption or the file's metadata.

Section 2 · Keyboard shortcuts

Keyboard shortcuts

The full set of shortcuts for driving idXpresso from the keyboard, grouped by context.

Command Shift Option Return Delete Escape

Windows & Modes

Face Recognition (reopen main window)R
Identity ManagerI
Learning Mode · ProL
Add Identity (Quick Add)N
Settings / Preferences,

File

Open photo…O
Remove photo from queue
Clear queue

Faces

Next face
Previous face
Ignore other face(s)-
Confirm auto match(es)+
Confirm auto match(es) & commit
Confirm & commit without metadata

Photo

Next in queue
Previous in queue
Commit · Pro
Show caption (toggle)I
Write metadata on commit (toggle)I
Copy confirmed name(s) with title(s)C

View

Compact mode (toggle face list)L
Auto face zoom (toggle)T

Edit

Find duplicates…D

Face list · no modifiers

Active only when the face list has focus and no text field is being edited.

Confirm selected face
Open identity pickerSpace
Ignore selected faceI
Reject selected faceX
Delete / reject selected face

Identity picker popover

Select picker item 1–9⌘1⌘9
Select picker item 100
Highlight next item
Highlight previous item
Close / cancel picker

Identity Manager · list pane

Select all visible identitiesA
Clear search text, else clear selection

Dialogs & sheets

Confirm / default action
Cancel

Blocked while a modal is open

These beep instead of firing while a modal is showing.

QuitQ
Close windowW
Section 4 · Settings

Settings

Everything in the Settings window, tab by tab. The tag on each row tells you the control type — toggle, picker, slider, and so on — so you know how to change it.

Open Settings with , (idXpresso ▸ Settings…). Pick a tab below to see its options.
Appearance
ModePicker
Sets the app's appearance.
SystemLightDarkBlack
Animated empty stateToggle
Shows a twinkling symbol animation when no photo is loaded.
Workflow
Auto-advance to next face after actionToggle
Selects the next unresolved face after you confirm, ignore, assign, or mark not-a-face.
Auto-select first face when opening a photoToggle
Selects the first detected face so review can begin immediately.
Compact Mode
Status indicationPicker
How a face's status is shown in the compact list.
Bullet ColorText Color
Name typePicker
What text the compact list shows for a matched face.
NameName and Title
Web Search
Web Search URL TemplateField
The URL used for a web name lookup. Use the {query} token where the name goes.
Image Search URL TemplateField
The URL used for an image lookup, also with a {query} token.

This tab controls how confidently idXpresso auto-matches faces. Pick a preset in Matching Mode, or choose Advanced controls to tune the individual thresholds yourself.

Matching Mode
Matching ModeRadio
The main control. Presets trade off strictness vs. speed; the last option unlocks manual tuning below.
CarefulBalancedFast LabelAdvanced controls
Advanced / Effective settings
Editable only when Advanced controls is selected; otherwise these show the active preset's effective values (read-only).
Auto-Match StrategyRadio
How a match is decided.
Absolute threshold onlyRelative (winner dominance)Hybrid (recommended)
Relative strictnessSlider
How dominant the top candidate must be over the runner-up. 0–100%.
Absolute thresholdSlider
Minimum similarity to auto-match. 0–100% (default 85%).
Minimum qualitySlider
Rejects low-quality faces before matching. 0–100% (default 20%).
Safety floorSlider
A hard minimum no match may fall below, up to 100%.
Similarity floor (relative)Slider
Lowest similarity allowed for a relative-strategy match. 0–100% (default 60%).
Display & Filtering
Hide identities already assigned in this photoToggle
Removes already-used people from the suggestion list to avoid duplicates.
Sort identity suggestions byRadio
Order of the ID picker list.
Confidence (highest first)Name (A–Z)
Show Suggestion CountsToggle
Shows how many candidate matches a face has.
FormatRadio
Shown when counts are on.
Simple CountThreshold Breakdown
Suggestion Count ThresholdSlider
Confidence cutoff for the count. 0–100%.
Show face match scoreRadio
Whether (and how) to display the match score on a face.
Don't show scoresAbsolute similarityRelative confidenceAbsolute + relative
Embedding Settings
SourceRadio
Which faces contribute encodings.
All facesOnly small facesNo faces
Master
Enable Metadata WritingToggle
The master switch. When off, Commit still learns faces but writes no metadata to the file.
Caption
Name with TitleToggle
Prepends the caption with the identity name(s).
Use delimiter between names and captionToggle
Inserts an em-dash separator. Requires caption writing on.
Persons in Image
NameToggle
Writes names into the Persons-in-Image field.
Keywords
NameToggle
Adds each name as a keyword.
Name with TitleToggle
Adds name-with-title as a keyword.
Post-Commit Metadata
After commitPicker
What happens to metadata fields after a commit.
No changeSet class valueClear fields
Class valuePicker
Shown when "Set class value" is chosen. Class 0–8.
Non-Writable Files
Default actionPicker
What to do when a file can't be written to.
Always AskCopy CaptionCopy Names & Titles
Matching
Enable prefix matchingToggle
Lets partial names match — e.g. "Don" matches "Donald".
Synonyms
Enable synonymsToggle
Treats configured alternate terms as equivalent in search.
Custom synonyms (JSON)Editor
A text editor for your synonym map. Format: { "term": ["synonym1", "synonym2"] }.
Apply · Revert to DefaultsAction
Save your synonym edits, or reset them to the defaults.
Cache Limits (MB)
TotalField
Overall thumbnail cache cap. 100–5000 MB, step 50.
128px tierField
Cap for small thumbnails. 50–4000 MB, step 50.
512px tierField
Cap for large thumbnails. 50–4000 MB, step 50.
Retention
Delete unused afterField
Days to keep unused cache before pruning. 1–365.
Clear cache on quitToggle
Empties the thumbnail cache each time you quit.
Actions
Refresh · Clear Thumbnails · Rebuild IndexAction
One-off maintenance — these aren't saved settings.

This tab is for management and maintenance — it shows information and offers actions rather than toggles you set.

Database Location
Change Location…Action
Move the database to a folder you choose.
Revert to Default LocationAction
Shown only when a custom location is set — moves it back.
Information
File path · Size · Last modified · StatisticsRead-only
Details about the current database.
Maintenance
Delete Encoding(s) · Repair Encoding Status · Remove All Photo Links · Delete DatabaseAction
Housekeeping and destructive actions — use with care.
Watch Folder · Pro
Enable Watch FolderToggle
Turns on automatic processing of a folder as new photos arrive.
Choose Folder…Action
Pick the folder to watch (with a Clear Selection option).
Process subfoldersToggle
Also watches nested folders (recursive).
Use metadata trigger gateToggle
Only process files that carry a matching metadata marker.
Target Color ClassPicker
Shown when the gate is on. Class 0–8.
Import ModePicker
Which incoming files get processed.
Only NewNew + ChangedAllow Reprocess
Test Watch FolderAction
Runs a check to confirm the folder and rules work.
Background Activity
Enable background detectionToggle
Detects faces in the background. Takes effect on next launch.
ConcurrencyField
How many photos process in parallel. 1–4.
Folder Permissions
Add Folder… · Clear AllAction
Grant or revoke read access to folders that hold your local photos.
Diagnostics
Export Diagnostics… · Copy Build InfoAction
Export a diagnostics bundle or copy build details (with a read-only build summary shown).
App Lock
Enable App LockToggle
Require Touch ID or your passcode on launch.
Auto lock after N minutesToggle + Field
Locks after an idle period. 1–240 minutes.
Auto lock when leaving app windowToggle
Locks as soon as idXpresso loses focus.
Auto lock on sleepToggle
Locks when the Mac goes to sleep.
Secure Deletion
Secure cache thumbnail deletionToggle
Overwrites cached thumbnails when deleting them.
Secure database deletionToggle
Overwrites the database on delete. Takes effect on next launch.

Account and plan management rather than settings you toggle.

Plan
Current plan · Manage SubscriptionAccount
Shows your plan; Manage Subscription appears for subscription-backed plans.
Available PlansPurchase
Buy a plan (shown on Free / Expired).
Feature comparisonRead-only
What each plan includes.
Restore PurchasesAction
Re-applies a purchase made on another device or after reinstall.
Terms of Service · Privacy PolicyLinks
Opens the legal documents.
Section 5 · Custom views

Understanding custom views

idXpresso's View menu lets you tailor the workspace — strip distractions, zoom in on a face, or hide boxes as you resolve them. Flip the switches below and watch the window respond.

Where to find these: most live in the menu bar under View. A couple are set in Settings. The switches here are just a preview — they don't change any real setting.
‹ 1/1 ›
Face 1Ignored
Face 2 ⌄
John Doe ◆ SharpestCOO John Doe
Jane Doe (88%)Auto-matched
Face 5 ⌄
Commit
Flip a switch — the window updates live.

From the View menu

Status indication — in compact mode, show a face's status as a colored bullet or as colored name text. Set in Settings ▸ General.

What the sharpest badge looks like

It's a small yellow corner tag — on the face's box in the photo, and on its thumbnail in the list (next to the name when you're in compact mode).

In the photo — the tag sits on the corner of the sharpest face's bounding box.
Face 3
In the list — the same tag appears on the row's thumbnail.

More view options in Settings ▸ Recognition

The Recognition tab has several controls that change what the ID picker and faces show:

1

Sort identity suggestions by

Order the ID picker list by confidence (highest first) or by name (A–Z).

2

Show suggestion counts

Show how many candidate matches a face has — as a simple count or a threshold breakdown, with an adjustable confidence cutoff.

3

Show face match score

Choose whether to display a score on a face, and which: absolute similarity, relative confidence, both, or none.

4

Hide identities already assigned

Removes people you've already used in this photo from the suggestion list, so you don't pick a duplicate.

Section 6 · Watch folder

Understanding the Watch Folder

Point idXpresso at a folder and it watches for new photos — dropping each supported image straight into your queue, no manual importing.

The watch folder lets you define a folder that idXpresso keeps an eye on. When a new supported photo appears there, the app automatically adds it to the queue. You enable it and tune how it behaves in Settings ▸ Watch Folder — choose the folder, watch subfolders, and gate imports by a metadata color class.

Here's why it matters: if your editing or cataloging app supports color classes (labels), you can effectively "send" a photo into idXpresso just by changing its color class in that app. idXpresso sees the change and pulls the photo in — which is what lets it slot neatly into an existing professional editing workflow instead of replacing it.

You also control it right from the main window. In recognition mode, click the watch-folder icon in the bottom toolbar to enable or disable it and jump to its preferences — the quick controls pop up right there.

1/1
Face 1Ignored
Face 2
John Doe (90%)
COO John Doe
Face 4
Face 5
Commit
A
Watch folder icon
Detail A · the watch-folder icon, magnified

Icon status key

The folder icon's color tells you the watch folder's state at a glance:

Off

The watch folder is disabled or no folder is set. Click the icon to turn it on.

Green — on

Set, active, and monitoring the folder for new photos.

Orange — error

Something needs attention — for example the watched folder is missing or unreadable.

Section 7 · Metadata writing

Understanding metadata writing

idXpresso finishes captions instead of replacing them — it reads the caption already on a photo, adds the confirmed names, and writes everything back to the standard fields your other photo apps already read.

When you open a photo, idXpresso automatically reads any existing IPTC/XMP caption and loads it into the caption field. This is built for a professional workflow: many teams pre-caption images with a general description the moment they ingest them, then let idXpresso complete the caption by adding the identified names to the front.

Nothing is written until you press Commit. At that point idXpresso writes to the proper IPTC/XMP fields — caption, keywords, and persons in image — all standard fields that leading photo-editing programs can read. You can turn writing on or off at any time with the Write Metadata toggle in caption mode, or set the default in Settings ▸ Metadata.

1

Existing caption, read in

1/1
Caption Write Metadata
…walk from a helicopter last week after landing at the airport.
Commit

The general caption was added when the photo was ingested. idXpresso reads it in automatically — faces are detected but not yet confirmed.

2

Confirm the faces

1/1
Jane Doe
test person1
CEO John Doe
test person2
Test person3
Commit

Confirm each face in the list. The boxes turn blue and the names are now ready to be written into the metadata.

3

Names written into the caption

1/1
Caption Write Metadata
Jane Doe, test person1, CEO John Doe, test person2, and Test person3 walk from a helicopter last week after landing at the airport.
Commit

On Commit, the confirmed names are added to the front of the existing caption — the original description is kept intact.

What you control

Everything below lives in Settings ▸ Metadata, so idXpresso fits the way your team already works.

Which fields get written

Turn each destination on or off independently. All three are standard fields that professional photo software reads.

CaptionKeywordsPersons in Image

How names join the caption

Add names to the beginning of the caption as usual, or separate the added IDs from your existing caption with an em-dash — an extra layer of safety that keeps automated work visibly distinct from your original writing.

Mark processed photos with a color class

idXpresso can stamp a committed photo with a color class — another standard piece of metadata pro software can see — so you can spot which photos are done, or drive a multi-app workflow from that label.

When a file can't be written

If a file's metadata can't be updated, idXpresso falls back to the clipboard — copying either the full caption or just names & titles. Set it to happen every time, or to ask each time.

idXpresso mascot presenting
Part 3

Workflows

The everyday, start-to-finish flows for captioning at speed — from import to committed metadata.

idXpresso is built to fit into a professional photographer's or editor's workflow. Because it reads and writes standard metadata, it drops into whatever tools you already use — there's no need to change your process. It's deliberately flexible, but it's designed with one job in mind: to finalize captions and metadata with specific identities, right before a photo is archived or submitted.

Your editor handlesthe what, where & when
+
idXpresso handlesthe who

Where it sits in the pipeline

A typical professional flow — idXpresso steps in near the end, once your selects are captioned and corrected.

Ingest OS or editor

Copy or move images from the camera or memory card onto your working drive.

Batch caption Photo editor

Apply a generalized caption across the shoot — the what, where & when — plus any other standard metadata.

Cull Photo editor

Review the shoot and select the keepers.

Color correct Photo editor

Edit, grade, and retouch the selected images.

Detect, match & caption idXpresso

Detect faces, match identities, and automatically complete the caption — the who — then write it to standard metadata.

Submit, transmit or archive Photo editor

Send the finished, fully-captioned images on to their destination.

Workflow 1 · First run

First run: build your database as you work

First run is a simple workflow to add identities and their faces as you work. We recommend folding it into your normal captioning routine — it's the easiest way to progressively build your recognition database.

On first run, idXpresso's database is empty. For facial recognition to work you have to train it — recognition needs both an identity and its face encodings. This workflow creates both.
1

Launch idXpresso

Recognition mode opens by default the first time you start the app.

  • Reopen it any time from File ▸ Face Recognition, or press ⇧⌘R.
⇧⌘R
Recognition
Drop a photo to get started
i No identities yet — detect faces and type names to get started
2

Open a photo

To process an image, open it into the recognition window.

  • Drag an image anywhere into the window, or press ⌘O.
  • Drop a supported file onto the app icon and idXpresso opens it automatically.
Recognition
Drop a photo to get started
i No identities yet — detect faces and type names to get started
3

idXpresso detects the faces

Wait a moment while idXpresso finds every face in the photo.

  • Once trained, this step also applies auto-matches and suggestions.
Recognition
i No identities yet — detect faces and type names to get started
Face 1
Face 2
Face 3
4

Go to a face

Jump to a face by clicking its bounding box in the photo, or by selecting it in the face list — the two stay in sync.

Recognition
i No identities yet — detect faces and type names to get started
Face 1
Face 2
Face 3
5

Click the face's name

Click the name — e.g. “Face 1.” Faces are numbered by position, left to right across the photo.

  • After training, auto-matched faces show the person's name instead of a number.
Recognition
i No identities yet — detect faces and type names to get started
Face 1
Face 2
Face 3
6

Name the face in the picker

When the picker opens, type the person's name. That opens the Quick Add dialog, where you enter the name and — following your own style — the name with title. (The title field takes the name again, so it reflects your exact formatting.)

  • The picker filters as you type once identities exist in the database.
  • Pre-loaded an identity with no encodings (via CSV import or the Identity Manager)? Select it here to link the face.
  • After training, the picker suggests matches for you.
Recognition
i No identities yet — detect faces and type names to get started
Face 1
Jane Doe
↩ Add “Jane Doe” as new identity
Quick Add
7

Check the caption

Open the caption viewer to see how idXpresso has added the names and titles to the beginning of the caption.

Caption
Caption
Jane Doe, John Doe walk from a helicopter last week after landing.
8

Commit — save & learn

Click Commit to write the metadata to the photo — and to commit the new faces to their identities.

  • Committing encodings is how idXpresso learns. It weighs each encoding's quality: the first one for an identity is always accepted, and as an identity gathers more, the app curates them to sharpen auto-matching and suggestions.
Recognition
Jane Doe
John Doe
Commit
Workflow 2 · Researcher

Researcher: batch-import identities with a spreadsheet

A more advanced way to train idXpresso — ideal before a big event, or whenever you want to build the database around the people you know you'll need. Prepare identities, biographic data, and even face encodings in bulk from a CSV.

This workflow uses CSV Import. It pre-populates the Identity Manager with names and biographic data so you can later find and assign faces — and, if you include a photo link, it can generate encodings to train recognition too.
1

Create the CSV

Make a spreadsheet with one to four columns, depending on how much data you want to import.

  • Column 1 (required): Name.
  • Column 2 (optional): Name with Title.
  • Column 3 (optional): Biographic link.
  • Column 4 (optional): Photo link — local or hosted, and suitable for recognition (a single, clear, well-lit, head-on face).
names.csv
Name *Name + TitleBio linkPhoto link
2

Populate & verify

Fill in your data and double-check it before importing — clean rows in mean clean identities out.

names.csv
Name *Name + TitleBio linkPhoto link
Jane DoeDirector Jane Doebio.co/jdjane.jpg
John DoeCEO John Doebio.co/jnjohn.jpg
A. RayDr. A. Raybio.co/array.jpg
3

Import the CSV

Bring the spreadsheet into idXpresso.

  • Drop a CSV onto the app while in recognition mode and idXpresso automatically switches to CSV Import.
  • Or start it from the File menu.
Recognition
Drop a photo or CSV
i No identities yet — detect faces and type names to get started
CSV
→ CSV Import
4

Map the columns & import

Use the mapping controls to assign each column to the right field. To build encodings from the photo links, turn on Encode URL(s), then click Import.

CSV Import
Map CSV Columns
Name *
Column A
Name with Title
Column B
Bio Link
Column C
Photo Link
Column D
First Row is Header Encode URL(s) Import
5

Your database is primed

Now when you process a photo, the ID picker and Identity Manager are both populated with your names.

  • If you included photo links and enabled Encode URL(s), auto-matching and suggestions become available too.
Identity Manager
Jane Doe
John Doe
A. Ray
M. Chen
Workflow 3 · Learner

Learner: batch-train from captioned photos

Another advanced way to train idXpresso — like CSV import, but driven by photos whose metadata already carries the identity. It's an excellent way to train even when you're offline.

This uses Learning Mode. Each photo needs a single, clear, head-on face and the subject's biographic info embedded in its IPTC/XMP caption.
1

Caption the photo

Take an appropriate photo of the subject — head-on, well lit, a single face — and put the biographic data in the caption using this format: NAME; NAME WITH TITLE.

jane.jpg
Caption (IPTC/XMP)Jane Doe; Director Jane Doe
2

Open Learning Mode

Open Learning Mode to start feeding it captioned photos.

  • From the menu, or press ⌘L.
⌘L
Learning Mode
Drop captioned photos
Records appear here
3

Drop the photos in

Drop your captioned photos into Learning Mode. idXpresso reads the metadata, automatically creates a record, and adds the encoding to it.

  • Prefix the caption with idxpresso: (e.g. idxpresso: Jane Doe; Director Jane Doe) and the app recognizes it as a learning photo and enters Learning Mode automatically.
Learning Mode
StatusFileNameName w/ Title
Valid jane.jpgJane DoeDirector Jane Doe
1 photo +1 learned
JPG
Workflow 4 · Poweruser

Poweruser: caption at keyboard speed

Once it's trained, idXpresso IDs faces and writes captions faster than any macro — ideal for large groups or the occasional forgotten face, all while it keeps perfecting your database. It's built to be fast: you can process a photo in a few keystrokes and never touch the mouse.

Set it up for speed

The speed comes from preparation. Get these three things in place and the day-to-day becomes almost frictionless.

Train your app

Keep it current and prepare for shoots before they happen.

Watch Folder trigger gate

Enable the Watch Folder with a color class as the metadata trigger gate — then "send" a photo straight from a third-party editor to idXpresso just by changing its color class.

Metadata writing

Turn on metadata writing and tune it to your archiving style. Enable post-commit color class to restamp a photo after commit — handing it off cleanly to your other apps.

The fast pass

A trained database turns each photo into a short burst of keystrokes — select, pick, confirm, commit, next.

Recognition
Jane Doe
John Doe
Face 3
Committed · next
⌘→ Space ⌘1 ⌘↩

Select → pick → confirm → commit — boxes turn blue and the photo commits, hands-free.

1

Open your photo(s)

Open a single image or a whole batch into idXpresso — the queue fills up ready to work.

2

Fly through the queue

Move between queued photos without reaching for the mouse — previous ⇧⌘← · next ⇧⌘→.

3

Work each photo — entirely from the keyboard

  • Move between faces — previous ⌘← · next ⌘→.
  • On a yellow box, open the ID picker with Space, then press + the number beside a name to select it instantly.
  • On a blue box (auto-match), confirm it with . Or accept everything at once: ⌘↩ commits the photo, and ⇧⌘↩ confirms all pending auto-matches and commits in a single stroke.
  • Inspect or edit the caption⌘I.
  • Commit⌘↩ — writes the metadata, learns the faces, and drops you onto the next image.