June 22, 2026 · 7 min read

How to Use Keywords in Lightroom Classic (Without Getting Overwhelmed)

I spend a lot of time in Lightroom's metadata layer, partly because I build tools that sit on top of it. Face Tagger applies person-name keywords using face recognition. Smart Search lets you find photos by typing a description, then optionally writes those descriptive tags as keywords. Both tools revealed the same pattern: photographers with an organized keyword hierarchy find things instantly; photographers without one search by date and hope.

Keywords are not the most exciting part of Lightroom Classic. But they are the part that makes the catalog pay off at scale. Here is how they work, how to add them without losing an afternoon, and where the math breaks down badly enough that AI tools are the better answer.

What are keywords in Lightroom Classic, and where do they actually live?

A keyword in Lightroom Classic is a text tag stored in the catalog database. It is not a folder name, not a collection, and not part of the filename. When you tag a photo "coastal-maine", that string lives in the catalog alongside the photo's develop settings and other metadata.

Keywords can be written to the photo file on request via Metadata > Save Metadata to File (Cmd+S on Mac, Ctrl+S on Windows). For raw files this creates an XMP sidecar. For JPEGs and TIFFs it writes into the file itself. Either way, the keywords travel with the file if you move it to another catalog or application that reads XMP.

The distinction from collections is worth stating clearly. A collection is a curated grouping, assembled manually or automatically. A keyword is a property of the photo itself. You can use keywords as the basis for a Smart Collection, so all photos tagged "client-wedding" appear in a Smart Collection filtered on that keyword automatically. But the keyword exists independently of any collection.

What are the four ways to add keywords in Lightroom Classic?

Lightroom Classic gives you four entry points, each suited to a different moment in the workflow:

1. Keywording panel (Library module, right panel). Select one or more photos, then type in the "Click here to add keywords" field. Separate multiple keywords with commas. This is the workhorse method for post-shoot tagging sessions.
2. Painter tool (Spray Can). In Grid view, click the Painter icon in the toolbar (press T to show the toolbar if it is hidden). Set the mode to Keywords, type your keyword, then click or drag over photos. Fast for applying one keyword to many scattered photos without selecting them first.
3. Import dialog. The Apply During Import panel on the right side includes a Keywords field. Good for tags that apply to the entire shoot: the shoot name, location, client name, or event. Those tags go on everything in the import automatically.
4. Batch select then type. Select all photos you want to tag (Cmd+A or Ctrl+A for all, or Shift-click a range), then type in the Keywording panel. The keyword lands on every selected photo at once. This is how I apply broad category keywords to a whole shoot in under a minute.

Two other panels are worth knowing about. Keyword Suggestions shows keywords you have used on recent similar photos. Keyword Sets lets you save a palette of nine frequently used keywords assignable by pressing number keys 1-9. For high-volume work, these cut typing time significantly.

Why do photographers skip keywording, and why does June 2026 change the math?

The honest answer is time. A 400-photo shoot with 50 unique subjects, locations, and moments requires hundreds of individual tagging decisions. Multiplied across a full year of shooting, that is days of catalog work that produces no new images and earns nothing. Photographers skip it because the payoff is deferred and the cost is immediate.

Two things change that calculus now.

First, Lightroom Classic 15.4 (released June 2026) added keyword sync to the cloud ecosystem for the first time. Keywords you apply in Lightroom Classic now propagate to Lightroom mobile and the web interface. An empty keyword list used to mean Classic could not find things. Now it means neither Classic nor mobile nor the web can find things. The blast radius is wider.

Second, AI tagging tools have gotten good enough to handle the repetitive parts. Face recognition can apply person-name keywords across thousands of photos without manual click-per-face work. Content-based search can identify photos by scene description without requiring pre-applied tags at all. The keyword problem has not disappeared; the tools that handle it have improved dramatically.

What keyword hierarchy actually holds up at scale?

The Keyword List panel in Lightroom Classic supports a full tree structure: parent keywords with nested children, any depth. The temptation is to build a taxonomy that would satisfy a professional librarian. The reality is that deep hierarchies become confusing and hard to maintain over time.

A flat top level with four to six categories holds up better in practice:

The test I apply before adding any keyword: "Would I realistically search for this?" If the answer is no, the keyword is noise. Keywords you never filter on slow the Keyword List and clutter the autocomplete suggestions when you start typing.

One structural rule worth following: keep People keywords specific and nested under a People parent rather than at the top level. This lets you filter on the parent to see all people-tagged photos at once, which is useful for checking coverage before delivering to clients.

Lightroom Classic can import a keyword list from a plain text file via Metadata > Import Keywords. One keyword per line; tab-indented lines become children of the parent keyword above them. Build your full hierarchy in a text editor once, import it, and you are done. No typing each entry individually into the panel.

AI auto-tagging: two flavors that solve different problems

Manual keywording breaks down in two specific ways. The first is people: recognizing a face and applying the correct name requires looking at every face in every photo. At 30,000 photos this stops being a workflow problem and becomes a combinatorics problem. The second is content: describing what is in a photo ("sunset over water", "child laughing", "empty chair by window") requires language that varies per image and is impossible to standardize into a fixed keyword palette.

These two problems call for different tools.

Face recognition for person keywords. Face Tagger runs a local ArcFace model against your library, clusters faces by identity, and lets you name the clusters. Once you name a person, it applies that person's keyword to every photo containing their face across the entire catalog, retroactively. A library with 5,000 photos of the same 12 people can be fully name-tagged in an afternoon rather than weeks of manual work.

Content-based search for scene keywords. Smart Search takes a different approach. It builds a semantic index of your photos using CLIP embeddings, then lets you search by typing what you are looking for: "sunset at the beach", "people laughing around a table", "foggy city street". It finds photos matching the description even if no keyword was ever applied. You can optionally write matching results as keywords, layering the Smart Search index directly on top of your keyword structure.

A third category worth knowing exists: microstock keywording tools like LrTag or Wordroom generate keyword suggestions optimized for stock marketplace vocabularies. These serve a different goal, maximum discoverability on a stock platform, and produce different output (40-50 generic keywords per photo rather than 5-15 specific ones). If you sell on microstock, the vocabulary and density requirements differ from personal library management, and a dedicated tool for that market makes more sense than a general tagging workflow.

Frequently asked questions about Lightroom Classic keywords

Do Lightroom Classic keywords write to the file or just the catalog?

By default, keywords live only in the catalog. To write them to the file, use Metadata > Save Metadata to File (Cmd+S on Mac, Ctrl+S on Windows). For raw files this creates an XMP sidecar. For JPEGs and TIFFs it writes into the file itself. Catalog Settings has an option to write metadata automatically on every change, which keeps files and catalog in sync without manual saves.

How many keywords should a photo have?

For a personal or editorial library, 5-15 specific keywords is a reasonable range. More than that and the value of individual keywords decreases; fewer and you may not have enough signal to filter effectively. Microstock workflows use 40-50 per photo because platforms weight keyword count in search rankings, but that logic does not apply to a personal catalog where you control what gets surfaced.

Do keywords from Lightroom Classic show up in Google Photos after you publish?

No. The Google Photos API does not accept keyword or tag metadata on upload, and Google Photos does not read XMP data from the image file. Your keywords stay in the Lightroom catalog and any XMP sidecars. They do not appear in the Google Photos interface after publishing, regardless of which publish plugin you use.

What is the difference between keywords and Smart Collections?

Keywords are metadata properties stored with the photo and portable via XMP. Smart Collections are saved filter queries that automatically include photos matching rules you define, including keyword rules. A Smart Collection filtered on the keyword "coastal-maine" is a live view that updates automatically whenever you apply or remove that keyword. The keyword is the data; the Smart Collection is a window onto it.

Can I import a keyword list from a text file into Lightroom Classic?

Yes. Go to Metadata > Import Keywords in the Library module. Lightroom Classic accepts a plain text file with one keyword per line. Tab-indented lines become children of the parent keyword above them. Build your hierarchy once in a text editor, import it, and every keyword is ready to apply without typing each entry into the panel one by one.