Skip to content
Learning Outcomes
  • Understand what tags are and how they are scoped.
  • Assign and remove tags on runs, projects, pipelines, and datasets.
  • Filter a list down to a tag.
  • (Admins) Manage the tag vocabulary for each type of item.

Organizing with Tags

A tag is a short, colored label you attach to an item so you can find it again later. Once your work grows past a handful of runs, names and dates stop being enough. Tags let you mark what something is (qc-passed, deprecated, reference) rather than what it is called.

You can tag four kinds of item:

Item Assign tags from Tags are shown on
Run The run's Settings The runs list, both inside a project and across projects
Project The project's Settings The projects list, and project cards on the Activity page
Pipeline The pipeline's Settings A project's pipelines list, and the pipeline glossary cards
Dataset The dataset's detail page The datasets list

Each Type of Item Has Its Own Tags

Tags do not cross between types. A qc-passed tag created for runs is offered only when you tag a run; it will never appear in the list of tags for a dataset, and the same name can exist independently for both.

Why they are kept separate

The vocabulary that makes sense for a dataset (raw, processed, reference) has little to do with the one that makes sense for a run (qc-passed, delivered). Keeping them separate means each dropdown stays short and relevant instead of becoming a single long list where most entries do not apply.

Assigning a Tag

  1. Open the item. For a run, project, or pipeline, go to its Settings. For a dataset, open the dataset itself.

  2. Find the Tags section. It sits near the bottom, under the heading Tags, with the note "Assign tags to categorize this item." If nothing is assigned yet it reads No tags assigned.

  3. Choose a tag. Click Add a tag... and pick one from the dropdown. Each entry shows the tag's color alongside its name. The tag is applied immediately, and appears as a colored chip.

  4. Remove a tag by clicking the × on its chip.

You can assign as many tags to an item as you like. A tag already on the item drops out of the Add a tag... list, so you cannot add it twice.

Nothing in the dropdown?

The list only offers tags that an administrator has created for that type of item, and only ones that are still active. If it is empty, ask an administrator to add some. See Managing the Tag Vocabulary below.

Who Can Assign Tags

Tagging an item requires edit access to that item, the same access you would need to change any of its other settings. If you can only view an item, you will still see its tags, but the Add a tag... control and the × buttons are hidden.

This means tags are collaborative: anyone who can edit a shared run can also adjust how it is labelled.

Pipelines are tagged per revision

A pipeline tag applies to the specific revision you tagged, not to every revision of that pipeline. This is deliberate, and it is what lets you mark one revision production while a newer one is still beta.

Filtering a List by Tag

Every list that displays a Tags column also has a tag filter above it, shown as a Select tag dropdown. Pick a tag and the list narrows to items carrying it. This works on the runs list, the projects list, a project's pipelines list, and the datasets list.

Filtering respects permissions: you only ever see items you already have access to, so filtering by a tag cannot reveal someone else's work.

Managing the Tag Vocabulary (Administrators)

Tags are created centrally so that a team shares one consistent vocabulary rather than each person inventing their own spelling of "QC passed".

Go to Admin → Tags. The page has one tab per type of item: Runs, Projects, Pipelines, and Datasets. Whichever tab is active is the vocabulary you are editing.

From here you can:

  • Create a tag. Click Create Type Tag, give it a name and a color.
  • Load a starter set. Click Load Predefined Type Tags to add a sensible default vocabulary in one step. It is a useful starting point on a new instance:
Type Predefined tags
Runs deliverable, high-priority, qc-failed, qc-passed, review-required, demo, benchmark, delivered
Projects active, archived, shared, template
Pipelines production, beta, deprecated, validated, example
Datasets raw, processed, reference, public
  • Deactivate a tag you no longer want people to use.

Deactivating is not deleting

A deactivated tag disappears from the Add a tag... dropdown, so nobody can apply it to anything new. It stays visible on items already carrying it, and history is preserved rather than rewritten.

This is usually what you want when retiring a label. If you deactivate a tag by mistake, reactivate it and it becomes assignable again.