Pet and animal keywording lives or dies on naming. Buyers in this category search by breed or species, so a file tagged dog, cute, outdoors never reaches them, while one that says Siamese cat does. Adobe allows up to 49 keywords per submission and weights the first ten most heavily, which is ample room to name the animal exactly and still answer the broad query underneath it.
Who licenses animal images and what they ask for
The buyers split into groups that use completely different vocabulary. Veterinary practices, pet food and pet insurance brands shop for health, care and companionship. Animal shelters and rescue charities shop for adoption. Conservation groups, museums and publishers shop for species accuracy. Consumer brands want lifestyle scenes where the animal is a prop rather than the subject.
Only two of those four care about breed, but they are the two that license the most tightly specified images, and they are the ones a vague keyword set loses first. A rescue charity looking for an adoptable mixed-breed terrier and a pet insurer looking for a pedigree Labrador will not accept each other's results, so a file that tries to answer both usually answers neither.
The breed is the keyword, and Adobe says so plainly
Adobe runs a dedicated dog and cat breeds mission, and its instruction is unusually blunt for an agency brief: title the submission using only the animal breed. The worked example it gives is Siamese Cat. No scene, no adjective, no setting. That is close to the opposite of the descriptive sentence most keywording tools produce by default, and it tells you how the category is indexed.
The same brief excludes mixed breeds and non-recognisable varieties from that particular mission, which is worth reading carefully rather than as a general ban. Mixed-breed animals sell perfectly well to shelters and rescue buyers. They simply cannot be labelled with a breed they are not, and claiming one is the fastest way to be the file buyers scroll past.
The asymmetry here is the thing to internalise. A missing breed costs you the searches you would have appeared in, which is an opportunity lost. A wrong breed puts your file into searches where buyers actively reject it, which teaches the agency's ranking that your image is the one people skip. On a platform that learns from click-through, a confident guess is worse than an honest omission.
The general-to-specific chain Adobe actually publishes
Adobe's keyword guidance asks for general and specific terms together, and the example it chose to explain the idea is itself an animal one: Animal, mammal, carnivore. That chain is the layer most pet files skip entirely, jumping from the breed straight to a mood word. It is also the layer that catches the broad, high-volume queries a breed term alone cannot reach. The wider set of rules Adobe enforces on keywords sits behind this one, including the instruction to use each keyword only once and to stay in a single language.
Adobe also suggests separating descriptive elements rather than compressing them, and again picks an animal to show it: White, fluffy, young animal, pup. Four separate entries, not one phrase. Contributors who write young white fluffy pup as a single keyword get one term that almost nobody types instead of four that plenty of people do.
A generic animal keyword set, rebuilt
The same photograph keyworded twice: a descriptive set that only reaches mood searches, and a rebuilt set that names the breed, adds Adobe's general-to-specific chain, then behaviour and buyer concept terms.
Wildlife: the binomial is a claim, not a decoration
Conservation and publishing buyers search scientific names, and they search them because the common name is ambiguous. A file carrying the correct binomial reaches a small audience that licenses reliably and rarely haggles. That makes the scientific name one of the highest-value single keywords in the whole category.
It is also the keyword with the least tolerance for a guess. A plausible but incorrect binomial does not simply miss; it asserts something false about the photograph to exactly the buyers equipped to notice. If you cannot confirm the species from your own notes or a reference you trust, the right move is to keyword to the level you are sure of, which is usually the genus or the common group, and stop there.
Both the common name and the scientific name belong in the list where you know them, because the two audiences do not overlap. Captive and wild are worth distinguishing too: a zoo photograph described as wild is a misrepresentation that editorial buyers in particular check, and the honest label costs you nothing with the buyers who wanted a captive subject anyway.
Behaviour is what separates two identical portraits
Once the breed is named, every contributor shooting that breed has the same first keyword. What differentiates files after that is what the animal is doing. Sleeping, yawning, stretching, grooming, eating, drinking, playing, begging, swimming, digging and alert-watching are all distinct searches, and they are the terms a buyer reaches for when the breed search returns three thousand results.
Adobe's breed mission points the same way, specifically mentioning animals eating or playing with toys among the shots it wants. Age is the other cheap differentiator the category rewards: puppy, kitten, juvenile, adult and senior are separate markets, and senior animals in particular are under-supplied relative to what veterinary and insurance buyers ask for.
The concepts buyers are really shopping for
A pet insurer is not searching for a dog. It is searching for the idea of a dog being cared for, and the words it types are closer to pet insurance, veterinary care, pet health and responsible ownership than to any breed. This is the layer where animal photography converts into commercial licensing, and it is the layer a vision model describing the frame will never produce, because the concept is not visible in the picture. The general principle behind choosing keywords buyers actually type applies here with unusual force.
The discipline is to keep these honest. Adoption belongs on a shelter-appropriate image, not on a pedigree studio portrait. Veterinary belongs where there is a clinical or care context, not on every dog in a field. Concept keywords are the easiest place in this category to drift from describing into wishing, and a concept the image cannot support behaves exactly like a wrong breed.
The first ten positions carry the ranking
Adobe states that keyword order is essential and that the first ten positions have the greatest influence on search ranking. For this category that settles the running order: the breed or species first, its obvious variants next, then the general-to-specific chain, then behaviour. Mood words and background detail go at the back if they go in at all. The mechanics of structuring a keyword field matter here because a two-word entry between commas is one phrase, not two terms, which is how labrador retriever ends up indexed without labrador or retriever standing alone.
The ceiling is not the same everywhere. Adobe takes up to 49 keywords per submission. 123RF asks for a minimum of seven and describes roughly thirty to fifty as the useful range, so hitting the stated minimum satisfies the rule and wastes the file. Because 123RF also locks the text once a file is accepted, a breed you got wrong there is permanent, which is an argument for resolving identification before the batch goes anywhere.
People, property and what your keywords may claim
Animal photographs are rarely only animals. A hand on a collar, an owner half in frame, a silhouette at the edge of the shot: Alamy's own guidance is that a model release is needed for commercial use wherever a person can recognise themselves, and it specifically includes body parts and silhouettes in that. An arm holding a kitten is a person in the picture.
Property follows the same logic and is where the category trips people up. The requirement attaches to anything identifiable that is copyrighted or trademarked, signed by its owner, which in practice means the branded collar, the logo on the food bowl, the patterned designer bed and the recognisable private house behind the dog. The agencies' guidance does not treat the animal itself as property needing a release, but everything placed around it may be, and attaching the right release at upload time is what decides whether the file can be licensed commercially at all.
This caps the text, not just the licence. Without a model release the file cannot be keyworded as though it were cleared for commercial use, so the concept layer has to retreat to what the paperwork supports. Editorial normally needs no release, but editorial keywords describe the actual event factually and cannot borrow the commercial concept vocabulary. Deciding which of the two a file is, before writing the keywords, saves rewriting them.
Pre-upload checks for pet and animal files
The checks that decide a pet or animal submission: breed or species confirmed rather than guessed, Adobe's general-to-specific chain present, strongest terms in the first ten positions, and the release paperwork that caps what the keywords may claim.
Words this category is not allowed to use
Brand names are the obvious one, and they hide in animal shots more than contributors expect. The collar, the harness, the food bowl, the branded toy and the pet carrier all carry marks, and the rule is about the word rather than the object: Adobe asks contributors to keep trademarks, brand names and personal information out of the submission entirely, even when the item is genuinely in the frame.
Pet names are personal information and do not belong in the fields either. Nor do health claims: hypoallergenic, healthy, non-shedding and similar terms assert something a photograph cannot establish, and they read to a reviewer as the keyword stuffing both Adobe and 123RF warn against. 123RF names its own forbidden categories plainly, including unrelated terms, event dates and spelling mistakes, and a misspelled breed is worse than a missing one because it occupies a slot and matches nothing.
Where tools help and where they stop
The division of labour in this category is unusually clean. Behaviour, concept and the long tail automate well, and most tools produce them competently: Xpiks and StockSubmitter give you fast bulk editing on the desktop, PhotoTag.ai generates sets from the image, and managed platforms like Wirestock handle submission in exchange for a share of royalties. Identification does not automate. A fuller comparison of the tools in this category goes through where each one fits, but the constraint is the same across all of them: no current vision model reliably tells a Weimaraner from a Vizsla, and none of them knows whether the animal in your frame was captive.
That makes bulk correction the feature that matters more than generation quality. Rastock AI generates titles, descriptions and keywords from the image, applies each agency's banned and mandatory keyword rules, and keeps a rollback if a batch comes out wrong, which is the part that matters when one wrong breed has propagated across two hundred files. It charges a subscription rather than a share of your royalties, with a 14-day free trial and no card required.
What this comes down to
Name the animal, then widen. The breed or species goes first because that is what buyers type, the general-to-specific chain goes next because it catches everything the breed term misses, behaviour and age differentiate you from the other files sharing that breed, and the concept layer is what turns an animal photograph into a commercial licence. Everything decorative can wait until the back of the list.
The one habit worth building is refusing to guess. In most categories a vague keyword is a small loss. Here, an identification you were not sure about is a claim, it reaches the buyers best equipped to catch it, and on at least one major agency it cannot be taken back. Keyword to the level you can defend, and let the hierarchy pick up the rest.
Frequently asked questions
How many keywords should a pet or animal stock photo have?
Adobe accepts up to 49 keywords per submission and 123RF describes roughly thirty to fifty as its useful range, so a well-built animal file usually lands between twenty and forty. The count matters less than the order: the breed or species belongs in the first few positions, because Adobe weights the first ten most heavily.
Should I use the breed name in stock photo keywords?
Yes, and it should normally come first. Buyers in this category search by breed, and Adobe's dog and cat breeds mission goes further by asking contributors to title the submission using only the breed, with Siamese Cat as its example. The one condition is accuracy: never label a mixed-breed animal with a breed it is not.
Do I need a release for a pet in a stock photo?
Agency release guidance addresses identifiable people and recognisable copyrighted or trademarked property rather than the animal itself. What does trigger a release is everything around it: an owner's hand or silhouette needs a model release for commercial use, and a branded collar, bowl or a recognisable private house needs permission from the owner.
Should I include scientific names for wildlife photos?
Include the binomial when you are certain of the species, because conservation and publishing buyers search it and license reliably. If you are not certain, keyword to the genus or common group instead. An incorrect scientific name asserts something false to precisely the buyers equipped to notice, which is worse than omitting it.
Why do mood words like cute and adorable not work?
Buyers rarely search them, and both Adobe and 123RF ask for relevant terms only, so mood adjectives consume keyword slots without adding reach. They are not forbidden, but they belong at the back of the list behind the breed, the general-to-specific chain, behaviour terms and the concepts commercial licensors actually type.
Can I use pet brand names or the animal's name in keywords?
No to both. Adobe asks contributors to keep trademarks, brand names and personal information out of the submission entirely, even when a branded collar or bowl is genuinely visible in the frame. A pet's own name counts as personal information and adds no search reach, since no buyer is looking for it.