How often the data updates

Crown Intelligence does not refresh on a single global clock. Each register re-runs on its own configured schedule, matched to how often its own publisher issues updates — several times a day for a fast-moving sanctions feed, weekly or monthly for a steadier registry, and as published for a register with no fixed release calendar.

Every record and every register carries its own "as of" date, so how current a figure is stays visible rather than assumed. The compiled screening index is reprocessed after each register run, which is what lets the "Last data load" figure above move independently of any single register's own last-run time.

This matches the refresh commitment in our proposed commercial terms (not yet final), and is covered in more technical depth in Methodology's refresh cadence section.

If you hold an export of a register, re-fetch it on that register's own cadence shown in the table above rather than on a fixed calendar interval of your own choosing — a weekly register does not need a nightly pull, and a six-hourly one should not be left for a week.

Use these fields to detect what changed since your last pull, instead of re-downloading and diffing an entire file every time:

• loaded_at on a record — when Crown last confirmed that record against its source.
• first_seen on a record — when Crown first observed that record, useful for spotting additions since your last export.
• last_run_at on a register — the source register's own record of when that register was last re-run, independent of any single record inside it.

The source register itself is the machine-readable index behind this page: query GET /api/sources?q=<name> for any register's current cadence, last run and row count before deciding whether a re-fetch is due.

See the endpoint reference, keeping bulk data fresh, and how downloads work for bulk export options.