A directory or archive file is imported under a named dataset, purpose and handling rule, preserved byte-for-byte, and published as typed records that stay citable.

An existing approved extraction package is imported whole, its artefacts preserved, and its explicit records mapped into the canonical families without inventing conclusions.

A scoped endpoint and a versioned connector profile produce incremental acquisitions that keep their windows, source identifiers and change tokens.

Approved API responses and events are acquired with a raw-preservation policy and the source schema that described them, so a normalised record can always be traced back.

An authenticated gateway admits approved source events, makes them durable first, and only then normalises and indexes what is eligible.

Historical stores and exports gain proper source and acquisition context, keep their originals, and yield canonical records that can be improved over time.

A single identifier-and-time-range query reaches archives, repositories, device packages and feeds, and returns only authorised matches with exact evidence references.

A parser or schema update starts a bounded new transformation campaign whose outputs coexist with the retained outputs and citations that came before.

Signed local components run the full chain, from import through processing to search, inside a network boundary you establish and verify yourself.