Architecture is the one stock niche where the keyword that makes an image findable is often the same keyword that limits how it can be licensed. Naming a landmark, a named building or its architect is what gets you into the searches that matter, and it is also what can push the file into editorial territory. Getting this right is mostly about deciding which bucket a frame belongs in before you write a word.
Decide the bucket before you write the keywords
Property is releasable the way people are. A private house, a distinctive interior, a recognisable commercial building shot from private land: these generally need a property release to sell commercially, and without one the image goes out as editorial. That distinction is a licensing fact about the frame, not a stylistic choice, and it should be settled at ingest rather than discovered at submission.
The keywording consequence is direct. An editorial file wants the specifics: the building name, the city, the street, the date, the architect where relevant. A commercial file wants the opposite: generic, descriptive, unattributable language that lets a buyer use the image to represent a category rather than a place. Writing generic keywords on an editorial frame wastes its value. Writing specific ones on a commercial frame invites a rejection.
Keywording the editorial frame: be specific, and be right
Editorial architecture sells on precision. Buyers are illustrating an article about a city, a development, a building or a period, and they search accordingly. Work outward in layers: the building or structure, then the neighbourhood or district, then the city, then the region and country. A buyer looking for a shot of a specific district will not find you under the country alone, and a buyer covering the country will still reach you through the wider layers.
Accuracy is not optional here, because editorial images end up in journalism and educational publishing where a wrong building name propagates. Alamy in particular warns contributors that incorrect names of people and places have caused problems when buyers licensed images on the strength of a caption that turned out to be wrong. If you are not certain which building you photographed, find out before you caption it or describe it generically instead.
Naming the architect is one of the few places where a personal name is welcome in stock metadata. Alamy cites exactly this case as a legitimate use of personal data: the architect name attached to their building, where there is a clear benefit to the buyer. It is also a high-intent search term, because researchers and publishers look for work by practice rather than by city. Use it where you are sure of the attribution and leave it out where you are guessing.
Keywording the commercial frame: describe the type, not the address
Released property and generic construction are where the commercial money sits, and buyers search them as categories. Property type comes first: apartment, townhouse, detached house, office building, warehouse, retail unit. Then the era and style, since mid-century, brutalist, Victorian and contemporary are all real search terms. Then materials and structural features, because concrete, glass facade, exposed brick and timber frame all carry intent.
Interiors need the room named and the mood described. Kitchen, bathroom, home office and open plan living are obvious; less obvious is how much of the demand is emotional rather than architectural. Cosy, minimal, cluttered, bright, spacious, empty: buyers searching for an interior are usually illustrating a feeling about home or work, and those adjectives do more selling than the square footage ever will.
The concept layer is where property images actually get bought
A large share of real estate imagery is licensed by people writing about money rather than about buildings. Mortgage, rent, housing market, first-time buyer, downsizing, home insurance, property investment, construction industry: these are the terms attached to the articles that use your photographs. A clean shot of a suburban street is an illustration of the housing market long before it is an illustration of that street.
Layout terms belong here too. Architecture produces a lot of clean sky and blank facade, which is exactly what a designer with a headline to place is hunting for. Copy space is one of the highest-intent tags in the whole niche, and terms describing the angle and format, such as aerial view, street level, wide angle and vertical, narrow the result set in the way buyers actually filter.
What to strip out of a commercial file
Streets are full of readable text, and an AI keywording pass will pick up all of it. Shop signage, estate agent boards, vehicle liveries, street numbers and the visible names of businesses and their brands will land in your keyword list if nothing stops them. On a commercial file every one of those is a problem, and they arrive in batches rather than singly, which is why catching them by eye does not scale.
Private addresses are the other thing to keep out. A full street address on a private home is personal data about whoever lives there, and agencies restrict that in metadata regardless of how useful it might be for search. Name the district or the city instead. The buyer looking for that kind of image is illustrating a type of place, not a specific doorway.
The near-duplicate problem
A property shoot produces twenty frames of the same kitchen from angles that differ by a metre. Applying one keyword set across all of them is efficient and slightly self-defeating, because your own images then compete with each other for the same search and the buyer sees a wall of near-identical thumbnails. Cull to the frames that genuinely differ, and where you keep several, let the keywords say what is different about each: the angle, the light, whether a person is present, what is in the foreground.
The same shoot, four different keyword shapes
Agencies want this material described differently. Alamy expects a descriptive caption stating what the building is and where, with tags separate and up to ten of them promoted as supertags. Adobe Stock and Shutterstock want a compact title and a flat keyword list ordered by importance. Getty applies its own controlled vocabulary. Shipping one identical keyword set to all of them means at least one agency receives the wrong shape. The tooling side of this niche is covered in best tools for architecture and real estate stock content.
A pass over one image
- Decide editorial or commercial first, from the release and from what is identifiable in the frame. Everything else follows from that answer.
- Name the structure type, then the style and period, then the materials. Three layers before you touch anything abstract.
- Add location only at a level the licence supports: full specifics on editorial, district or city at most on commercial, never a private address.
- Add the concept layer that matches how the image will be used: housing market, urban development, home ownership, construction.
- Add the layout terms last, including copy space and the viewpoint, then read the list back looking for any brand or business name that slipped in.
Doing this on a full shoot
Two hundred frames of one property is a normal day, and the discipline above falls apart when it has to be applied by hand two hundred times. Rastock AI generates the metadata and then enforces your rules on it: a banned list that strips business and brand names before delivery, a mandatory list for the terms every file in a set must carry, and a rollback when a batch comes out wrong. It delivers to 10+ agencies over FTP, takes no revenue share and leaves your IPTC and CSV exports open. See the pipeline in detail or start the 14-day free trial, no card needed.
Frequently asked questions
Do I need a property release to sell architecture photos?
For commercial licensing, generally yes where the property is private or recognisable and identifiable as a specific place. Without a release the image can usually still be sold editorially. Because that decision changes how you keyword the file, settle it at ingest rather than at submission, and keep released and unreleased frames in separate folders from the start.
Can I put the building name in my keywords?
On an editorial file, yes, and you should: the building name is often the highest-intent search term you have. On a commercial file a specific named landmark can restrict how the image may be licensed, so describe the structure generically instead. Whichever route you take, only use a name you have actually verified.
Should I name the architect in my metadata?
Where you are certain of the attribution, yes. It is a strong search term for publishers and researchers who look for work by practice rather than by city. Alamy cites naming an architect alongside their building as a legitimate use of a personal name in metadata, on the grounds of clear benefit. Leave it out entirely if you are working from a guess.
What keywords sell real estate stock photos?
Much of this imagery is licensed by people writing about money rather than buildings, so concept terms carry real weight: mortgage, rent, housing market, first-time buyer, property investment, construction industry. Pair those with the property type, the style and period, and layout terms such as copy space, which art directors use to filter for images they can place text over.
Can I include a street address in the keywords?
Not for a private home. A full address is personal data about the residents, and agencies restrict that in metadata regardless of how useful it might be for search. Name the district or the city instead. Buyers searching this kind of image want a type of place rather than a specific doorway, so the broader term usually performs better anyway.
How do I keyword twenty near-identical shots of the same room?
Cull first, then differentiate. Applying one keyword set across all twenty makes your own images compete for the same search and shows the buyer a wall of near-identical thumbnails. Keep the frames that genuinely differ and let the keywords say how: the angle, the light, whether a person is present, what sits in the foreground.