Generation stopped being the hard part a while ago. A season of Flat-to-Model runs, a few packshot batches and a Model Swap set, and the library that used to hold a dozen flat-lays now holds thousands of images. The question is no longer "can we produce this". It's "which of these did the brand team approve, and which are still waiting on a retouch".

Tags are live today. Create a colored label once, put it on any generated result, uploaded asset or saved identity, and filter your gallery down to just those images. Tags work on every plan and cost no credits.

What a tag is

A tag is two things: a name of up to 64 characters, and a color. Ready, Needs retouch, SS26 campaign, Amazon are all fine, and so is anything else your team agrees on.

The important part is that there is one tag library, and it covers three kinds of thing: the images a job produced, the assets you uploaded, and the identities you generate with. So a single SS26 campaign tag can sit on the flat-lays you uploaded, the model you shot them on, and the results that came back.

To apply one, hover an image and click Tag. Start typing to search your existing tags, click any of them to add or remove it, or create a new one on the spot.

The tag popover open on an image in On-Model, showing a search field, a list of existing tags with color swatches, and the option to create a new tag — Search your existing tags, or create a new one without leaving the image.

This is a different thing from the Type labels and the AI search we shipped in Your Asset Library, Now Smarter. Those are detected automatically and describe what is in an image: a shirt, a bag, a person. Tags are written by you and record what you decided about it. The two work side by side, and you can filter on both at once.

Color-coded at a glance

Every tag carries a color, and tagged images show it as a small colored dot on the thumbnail. Hover a dot to see the tag name. An image with two tags gets two dots. The dots show up wherever your images do: your generated results, your uploaded assets, and your identities.

That sounds like a small thing. In a grid of two hundred images it's the difference between reading filenames and just seeing that the whole bottom row is still unapproved.

As you type a new tag name, On-Model suggests a color, picked consistently from the same name every time, so Ready lands on the same green for everyone who creates it. Eight suggestions cover most needs, and the swatch is a full color picker, so you can match an existing internal convention exactly.

The Assets page in On-Model, a grid of uploaded product images with small colored dots marking the tags on each one — The same dots on the Assets page. Two dots means two tags.

Narrow the gallery to one tag

Marking images only pays off if you can pull them back out. In the gallery, open Advanced Filters and use the Tags control to pick one tag or several. Selecting more than one is inclusive: you get every image carrying any of the tags you picked, which is what you want when you ask for "everything from the two autumn drops".

The tag filter stacks with the filters already there, so you can ask for shared work, from Flat-to-Model jobs, that finished successfully, and is tagged Needs retouch. Your selection survives navigation, so you can open a result, come back, and still be looking at the same filtered set.

The Advanced Filters panel in the On-Model gallery with the Tags multi-select open, showing selected tags with their color dots — Filter by one tag or several, on top of job type, status and ownership.

The same filter is on your assets, your identities, and the pickers you use when setting up a job. Tagging your five approved autumn models once means picking them from a five-item list next time, instead of scrolling your whole roster.

Your team already sees your tags

This is the part worth knowing, because nothing in the interface announces it: tags are shared across your group automatically. There is no share step and no permission to grant. If you and a colleague are in the same team, the tags you create appear in their tag list and their filters, and theirs appear in yours.

That makes a tag a piece of shared vocabulary rather than a personal bookmark. When a producer marks forty images Ready for PDP, the person handling the upload opens the gallery, picks that tag, and sees exactly those forty. Nobody exports a list, and nobody sends a spreadsheet of filenames.

The flip side is worth stating plainly: your teammates see your tag names and who created them. They are labels for collaborators to read, not private notes.

You can rename or recolor any tag you created, and the change follows it everywhere it has been applied. Teammates' tags stay theirs to edit. On top of both, On-Model ships a set of system tags everyone can use.

A vocabulary worth agreeing on

Tags reward a little upfront agreement. Three patterns cover most of what teams do with them:

Review state. Ready, Needs retouch, Approved, Rejected. The most common use, and the one that replaces the most chat messages.
Campaign or season. Winter27, SS26 campaign, Holiday drop. Cuts a year-round library back to the shoot you're working on.
Destination. Amazon, Zalando, Paid social, PDP. Useful when the same garment ships to channels with different requirements.

Two or three short lists beat one long one.

Available via API

Tags are in the API as well as the app, which matters when On-Model sits inside a larger pipeline: your DAM can mark what it ingested, or your review tool can write approvals back. Assign tags to an asset, creating any that don't exist yet, in one call:

PUT /assets/:asset_id/tags

import requests

api_url = "https://v2.api.piktid.com"

response = requests.put(
  f"{api_url}/assets/{asset_id}/tags",
  headers={"Authorization": "Bearer " + access_token},
  json={
      "tag_ids": ["4f1c...", "9ab2..."],
      "new_tags": [{"name": "SS26 campaign", "color": "#2E5A3B"}],
  },
)

The upload endpoint takes the same two fields, so files can arrive already labelled. Reading works the other way: pass tag_ids to the assets, identities, job results and gallery endpoints to get back only what carries those tags. Full reference is in the Tags API docs.

Common questions

What is image tagging in a fashion asset library? It's attaching your own short labels to images so you can retrieve them later by meaning rather than by filename. In On-Model a tag is a name plus a color, it can go on generated results, uploaded assets and saved identities, and every listing screen can filter by it.

Do tags cost credits? No. Tags are metadata. Creating, applying and filtering by them is free, on every plan, and applying one never changes how a job is processed.

Can my team see my tags? Yes, automatically. Everyone in your group sees the tags you create and can apply them to their own images, with no share step. You keep control of renaming and deleting the ones you created.

What happens to my tags when I regenerate an image? Regenerating or retouching an image produces a new version, and that new version starts untagged. Tags belong to the version they were applied to, so a re-run is a good moment to re-apply the review state.

Is there a limit on how many tags I can create? No. There's no cap on tags per account or per image, tag names don't have to be unique, and you can delete the ones you stop using.

Can I use tags through the API? Yes. There are endpoints to list, create, rename and delete tags, to assign and remove them on assets, identities and job outputs, and to filter every listing endpoint by tag.

Try it now

Open On-Model, hover any image in your gallery, and click Tag. Make one label for what you're working on this week and one for what's approved, then open Advanced Filters and watch the grid narrow.

New to On-Model? Start with the Flat-to-Model guide, the Create-Packshot guide, or the Model-Swap guide.