Photo metadata: the fields that pay you back
Photo metadata is three different things wearing one name, and only one of them is your job to fill in. This guide separates them, names the exact IPTC fields worth your time, shows you where they live in Lightroom Classic and Bridge, and tells you honestly which ones survive the trip to the internet and which ones do not.
Three kinds of metadata, and only one of them is your job
Every photo file carries information about itself. Some of it the camera writes without asking you, and some only exists if you type it. Knowing which is which clears up most of the confusion around this subject.
Exif is the camera's part. Shutter speed, aperture, ISO, lens, focal length, the moment the shutter fired, and on some bodies the GPS position. You do not enter Exif, you read it. It is what lets your software sort a two card wedding into one correct timeline.
IPTC is your part. The IPTC Photo Metadata Standard is the vocabulary that news agencies, picture libraries, museums and stock platforms all read. It holds who made the photo, how to reach them, what the photo shows, what it may be used for, and who owns it. The current version is 2025.1, published in November 2025, which added four properties for describing AI generated content.
XMP is not a third set of fields. It is how the other two get stored. XMP is Adobe's container format, now an ISO standard, and it is what Lightroom, Bridge, Photoshop and Capture One write into. For JPEGs and DNGs the XMP sits inside the file. For proprietary raw files it does not. Adobe's own documentation is blunt about why: "In the case of camera raw files that have a proprietary file format, XMP isn't written into the original files. XMP metadata is stored in a separate sidecar file to avoid file corruption."
That last point catches people out every year. If you fill in copyright on 800 CR3 or NEF files and then move the raws to a new drive without the matching .xmp sidecars, you have moved the photos and left the metadata behind. Our guide to managing raw files covers how to keep the pair together.
The one line version
Exif is what the camera recorded. IPTC is what you decide. XMP is the envelope both travel in.
The four fields that earn their keep
IPTC's own testing of how websites handle photo metadata is built around what it calls the 4Cs: caption, creator, copyright notice and credit line. If you never fill in anything else, fill in those four. Everything further down this page is a refinement on top of them.
| IPTC field | Stored as | What to put in it |
|---|---|---|
| Creator | dc:creator | Your name, or your studio name if you would rather not be named personally |
| Description | dc:description | The caption. Who, what, where, when, in plain sentences |
| Copyright Notice | dc:rights | Your copyright line, for example "© 2026 Jane Ruiz. All rights reserved." |
| Credit Line | photoshop:Credit | How you want to be credited in print, which is often not identical to your legal name |
Field names and XMP properties checked August 2026 against the IPTC Photo Metadata Standard 2025.1.
The reason these four beat everything else is that they travel inside the file. A watermark does not survive a crop or a screenshot. A line in your contract does not follow the JPEG into somebody's content management system. The Creator and Copyright Notice fields do, right up until a platform strips them, which is a real problem covered further down.
Copyright and contact: the fields that come back to you
These are the ones with a genuine payback, because they turn a stray copy of your photo into a lead or a license instead of a dead end.
- Copyright Notice (dc:rights). IPTC defines this as "any necessary copyright notice for claiming the intellectual property for this photograph." One line, with the year and your name.
- Rights Usage Terms (xmpRights:UsageTerms). Free text describing what somebody may do with the image. "Licensed to the client for personal use. No commercial or editorial use without written permission" is a perfectly good entry.
- Web Statement of Rights (xmpRights:WebStatement). A URL pointing at your licensing or usage page. This is the single most useful field on the list, because it is the one a picture editor can click.
- Creator's Contact Info (Iptc4xmpCore:CreatorContactInfo). A structured field with slots for email, website, phone and postal address. IPTC describes it as providing "all necessary information to get in contact with the creator of this image."
- Creator's jobtitle (photoshop:AuthorsPosition). Minor, but it is one keystroke in a preset.
The IPTC Extension schema goes further, with Licensor (Iptc4xmpExt:Licensor), Copyright Owner (Iptc4xmpExt:CopyrightOwner) and Image Creator (Iptc4xmpExt:ImageCreator) as structured properties. Those matter if you syndicate or license through an agency. For a working wedding or portrait photographer, the five above are the whole job.
None of this is a legal shield on its own. What it does is make you findable. An editor who wants to pay you has about ten seconds of patience, and an embedded website URL is what fits inside ten seconds.
Caption, headline and title do three different jobs
People fill these in interchangeably and then wonder why their stock submissions get bounced. They are not synonyms.
- Description (dc:description) is the caption: "a textual description, including captions, of the image." This is where full sentences go.
- Headline (photoshop:Headline) is "a brief synopsis of the caption." IPTC adds, in the standard itself, that "Headline is not the same as Title."
- Title (dc:title) is a short name for the image, closer to a label than a sentence.
Two newer fields matter if your photos end up on websites: Alt Text (Accessibility) (Iptc4xmpCore:AltTextAccessibility) for a short screen reader description, and Extended Description (Accessibility) (Iptc4xmpCore:ExtDescrAccessibility) for a longer one. Support across editing software is still uneven, so check that your app writes them before building a habit around them.
Three more fields quietly solve studio problems. Description Writer (photoshop:CaptionWriter) records who wrote the caption. Instructions (photoshop:Instructions) carries a note from you to whoever receives the file, which is the right place for an embargo date or a "do not crop" request. Job Id (photoshop:TransmissionReference) holds your job number, which is how you match a stray file back to an invoice two years later. If you work with a second shooter, agreeing on the Job Id format is worth more than any other single line in your handover notes, as our second shooter standards guide gets into.
Keywords: one field, and the one people overdo
Keywords live in a single IPTC field (dc:subject), defined as "keywords to express the subject and other aspects of the content of the image." It is a flat bag of words attached to the file, and everything clever your software does with hierarchies, parents and synonyms is a layer sitting on top of that one flat list.
Keywording is where most photographers spend the bulk of their metadata effort and get the least back, usually because they try to build a 400 term vocabulary in one sitting and abandon it by week three. It is genuinely valuable if you license images, pitch editors, or build blog posts out of your archive. It is close to worthless if you deliver client galleries and never search across shoots.
Because it needs a different kind of thinking to the fields above, we split it out: our guide to keywording photos covers building a small controlled vocabulary, using hierarchies without drowning in them, and how to tell when you should skip keywording altogether.
Enter it once, apply it forever
Nobody types a copyright notice 800 times. You build a preset or a template once, and then it lands on every file at import.
Lightroom Classic
In the Metadata panel, open the Preset menu and choose Edit Presets, fill in your standard fields, then choose Preset > Save Current Settings as New Preset. From then on you can apply it in the Import dialog and every incoming file arrives already stamped.
Three commands do the rest of the work. Metadata > Sync Metadata pushes the fields from one selected photo across a whole selection. Metadata > Save Metadata to File (Ctrl+S on Windows, Command+S on Mac) writes what is in the catalog out to the file or its sidecar. Metadata > Read Metadata From File pulls changes back in if you edited them elsewhere. If you would rather not think about it, turn on Catalog Settings > Metadata > Automatically Write Changes Into XMP and Lightroom keeps the files current for you, at a small cost in disk activity.
The Metadata panel has several views, including Default, EXIF and IPTC, IPTC, IPTC Extension, Location and Quick Describe. Set it to IPTC while you work on this, or you will spend the evening scrolling.
Adobe Bridge
Bridge is a free download and it is the better tool if you do not keep everything in a Lightroom catalog. Select files and choose File > File Info to edit fields directly. To build a reusable stamp, choose Tools > Create Metadata Template, tick the values you want, and save it. Applying it offers two modes from the Metadata panel menu: Append Metadata fills only the fields that are currently empty, and Replace Metadata overwrites what is there. Append is almost always what you want. Adobe notes that templates are "stored in a shared location that all XMP-enabled software can access," so a template you make in Bridge is visible to the rest of your Adobe apps.
Whichever route you take, do it at ingest. Metadata applied at the point of import survives a busy season. Metadata you promise to add on Sunday does not, which is one of the habits our ingest checklist is built around.
What survives the trip to the internet, and what does not
This is the part that deflates people, so it is better to know it early. IPTC has run a public test of how sharing platforms handle embedded photo metadata three times, most recently in spring 2019, with results published that August. The picture was not encouraging.
Instagram and Twitter stripped embedded metadata entirely, on both viewing and download. The free stock sites tested, Pexels, Pixabay and Unsplash, did the same. Flickr and Google Photos kept some metadata in the downloaded original, though Flickr substituted the account owner for the creator. Behance preserved it. Those results are seven years old and platforms change quietly, so treat them as a warning rather than current fact: upload one of your own files, download it back, and inspect it.
The practical conclusion is not that metadata is pointless. It is that metadata protects the files you control and the ones you hand over directly. Your archive, your delivered galleries, your licensing copies, the files you send an editor. Social platforms are a lossy channel and always have been, which is why the credit conversation there still happens in captions and DMs.
The one place embedded fields do pull weight publicly is your own website, where descriptions and keywords are part of what makes an image findable. That is a separate topic with its own rules, and we cover it in photo metadata for SEO.
Where this fits between the card and the client
The order that works is boring, and that is the point. Back up the cards to two places. Ingest, applying your metadata preset in the import dialog. Pick your selects. Edit the picks. Export, and check that your export preset is not stripping IPTC on the way out. Our guide to moving selections into Lightroom walks through the handoff, and the archiving guide covers what happens to the files after delivery.
One thing worth checking about any tool you add to that chain: whether it touches your files. Metadata you applied at import is only still there if nothing downstream rewrites, renames or relocates the file underneath it. Kepla clears the obvious misses and marks each one with a reason, and it never deletes, moves or renames a single file, so the fields you set at ingest stay exactly as you left them. The picking app for iPhone, iPad and Mac is still being built. The booking page is live and free today.
If you do nothing else after reading this, do these five things this week. Build one metadata preset with your name, copyright line, website and contact email. Apply it in the import dialog on your next job. Set your export preset to include IPTC. Open one delivered JPEG from last month and look at what is actually in it. Then leave it alone, because a metadata system you maintain badly is worse than four fields you fill in every single time.
FAQ
What is the difference between Exif and IPTC metadata?
Exif is written by the camera and records the technical facts of the exposure: shutter speed, aperture, ISO, lens, focal length, capture time and sometimes GPS. IPTC is written by you and records everything a camera cannot know: the caption, your name, your contact details, the copyright notice, usage terms and keywords. Both are usually stored in an XMP block, which is the container format rather than a third set of fields.
Where is copyright information stored in a photo file?
In the IPTC Copyright Notice field, stored as the XMP property dc:rights. Two related fields sit next to it: Rights Usage Terms, stored as xmpRights:UsageTerms, describes what somebody may do with the image, and Web Statement of Rights, stored as xmpRights:WebStatement, holds a link to your licensing page. Filling in all three takes about a minute once, inside a metadata preset that then applies to every import.
Why do my raw files have a separate .xmp file next to them?
Because Adobe will not write metadata into a proprietary raw format. Adobe's documentation says XMP is not written into camera raw files with a proprietary format, and is stored in a separate sidecar file to avoid file corruption. That sidecar holds your edits, ratings and IPTC fields. If you move or copy raw files, move the .xmp files with them, or you will arrive at the new drive with the photos and none of the work.
Does Instagram keep the copyright metadata in my photos?
IPTC's most recent public test, run in spring 2019, found that Instagram and Twitter stripped embedded photo metadata on both viewing and download. Platforms change their handling without announcing it, so the safest check is to upload one of your own files, download it back, and inspect what is left. Treat social platforms as a lossy channel and rely on embedded metadata for the files you deliver and license directly.
Do I need to fill in metadata on every photo, or just the ones I deliver?
Apply the standard fields, your name, contact details and copyright, to everything at import, because that costs you nothing once it is a preset. Apply the per image fields, caption, headline and keywords, only to the photos you actually selected. Captioning frames that will never leave your drive is the fastest way to abandon the habit entirely.
What is the 4Cs of photo metadata?
It is the shorthand IPTC uses in its own platform testing for the four fields that matter most: caption, creator, copyright notice and credit line. Between them they say what the photo shows, who made it, who owns it, and how the maker should be credited in print. If you are starting from nothing, fill in those four and add the contact and rights fields once the habit sticks.
Never sort photos at 1 AM again.
Kepla for Mac clears the obvious misses from a card, names the reason on every frame it sets aside, and leaves the choosing to you. Nothing is ever deleted, moved or renamed. Free through the private preview · the first hundred photographers keep it at $99 a year.