Data & Metadata

section-index · Reviewed 2026-11-02 GA

A job's results are only useful if you can find them again and know what they meant. Rescale gives you three ways to attach that context to a run, and which one you reach for depends on whether you are organizing your own work or holding a team to a standard.

Choosing a mechanism

Tags are the fastest thing to start with, because they need no setup: type a word onto a job and it becomes searchable. Custom fields take configuration in exchange for structure and enforcement, so they are what you use when the data has to be there rather than merely allowed. Automations sit behind both, filling either one from a script so a value lands without anyone remembering to type it.

You want to	Use	Because
Label and find things fast	Tags	Free-form, no setup, searchable with boolean logic
Enforce that data gets captured	Custom fields	Typed, scoped, and can be required before submission
Capture values without anyone typing them	Automations	A script extracts them from job output at runtime

They compose, and the combination is usually what a mature setup looks like: an automation extracts a result value, writes it into a custom field, and a tag marks the study it belongs to.

Tags

Free-form labels on jobs, workstations, and files. Boolean and wildcard search.

Custom fields

Typed, scoped key-value data, with required-field enforcement for governance.

Automations

Python or Bash scripts that populate metadata from job output automatically.

Tags or custom fields?

Most of the time this is the only question you need to answer, and the deciding difference is enforcement. A custom field can be made required, which blocks job submission until someone fills it in. A tag cannot, so a tag can always be skipped, which is fine when you are organizing and not when someone downstream is depending on the value.

	Tags	Custom fields
Shape	Unstructured text	Typed key-value pairs
Set up by	Anyone, on the spot	Anyone, on a job · Workspace Admins, workspace-wide
Can be required	❌	✅ — blocks submission until filled
Applies to	Jobs, workstations, files	Jobs and workstations, configured separately
If one is deleted	Gone with no record of who removed it	The field comes off existing work too, so export first
Best for	Finding things	Proving things

So: tags when you are organizing for yourself, custom fields when someone else is relying on the data being present.

What this enables

• Traceability — a complete picture of what was simulated, why, and what was concluded, for audits or production troubleshooting.
• Governance — required fields standardize what gets captured, so data quality doesn't depend on individual discipline.
• Analysis — consistent metadata across jobs turns a pile of runs into something you can query and compare.

Related

• Find and manage your jobs — the search syntax tags plug into
• Create and submit a job — where fields are filled in at setup
• Files and storage — tags apply to files too
• Billing & reports — custom fields appear as CSV columns