NERVE is live across local, private and approved-cloud deployments. Book an evaluation

Use cases

/

Archives and holdings

Turn an approved archive into searchable holdings.

Turn an approved archive into searchable holdings.

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.

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.

Audience

Archive and holdings teams, investigators who need to search material that arrives as a single delivery.

Source families

Files and archives, Documents and office content, Messages and email

The workflow

Source to searchable, step by step.

Source to searchable, step by step.

01

Input

Morgan receives an approved archive, either a nested directory tree or a container file, along with the dataset it belongs to, the purpose it was acquired for and the handling rule that applies to it.

02

Acquisition and preservation

NERVE records the source and acquisition context before anything is opened. Exact original bytes are retained and hashed. Container structure is expanded safely, and parent-and-child relationships are recorded so a message inside a mailbox inside an archive keeps its position in the delivery.

03

Transformation

Supported documents, messages and files are decoded, parsed and normalised into canonical records. Each record carries the common envelope, so a report from a shared drive and a message from a mailbox answer the same query fields.

04

Discovery and retrieval

Priya searches across the whole dataset by text, time, identifier and structure. Every result resolves back to the exact accepted bytes, and lineage shows which acquisition and which processing run produced it.

05

Governance

The dataset’s handling rule travels with every record produced from it. Search results are filtered by the searching account’s authorisation before counts and facets are built, so unauthorised material never shapes the view.

When it goes wrong

Failure is reported, not absorbed.

Failure is reported, not absorbed.

Failure variant

One corrupt child item inside the archive fails to decode. The job completes as Partial: the corrupt item is retained and marked, the failure is reported against that item, and every independent accepted item continues through to search.

What this does not claim

A hash match confirms byte identity. It does not confirm that two files mean the same thing, and it does not establish who created them.

Evaluation measures

Measure these on your own material.

Measure these on your own material.

Bring an approved sample and we will plan the run around these measures. The numbers belong to your environment, so they are established with you rather than quoted in advance.

Bring an approved sample and we will plan the run around these measures. The numbers belong to your environment, so they are established with you rather than quoted in advance.

Hash recovery for every retained original

Per-item visibility of warnings and failures

Canonical mapping coverage across the delivery

Exact citation resolution after reprocessing

Source-to-search time for the full dataset

Next step

See this workflow against your own holdings.

See this workflow against your own holdings.

An evaluation runs NERVE against a representative sample of your own sources, so you can measure acquisition, normalisation and retrieval behaviour on material you already understand.

An evaluation runs NERVE against a representative sample of your own sources, so you can measure acquisition, normalisation and retrieval behaviour on material you already understand.

Start an evaluation

All use cases

Architecture