Lines of investigation: records about people and things

What an analyst can lawfully find about a person, company, property or vehicle, where each record lives, and what Viokit does with it. Every connected source, with its grade, is listed in What each source is, and how far to trust it. Likelihood language follows the UK probability yardstick (Assessment language).

Status: built = Viokit looks it up; data pack = built, from an offline copy you install; next = planned, not built; catalogued = described, not connected; manual = record by hand (provenance kept); no = deliberately not built.

Ontology: the names Viokit uses

Viokit's vocabulary (22 kinds of thing, 38 relations) is its own, but it is mapped to the ontologies analysts already use (FollowTheMoney, used by OCCRP Aleph and OpenSanctions, and Maltego's standard entities), so a graph can move between tools.

Moving a case between tools

Three transfers (to FollowTheMoney, from FollowTheMoney, and to Maltego; people and agents alike) move a case between Viokit and those tools. None needs a network.

To Aleph or OpenSanctions tooling. Export writes a FollowTheMoney file, one entity per line. People, organisations, addresses, properties, vehicles, articles, court cases and events are entities. Links (directorships, ownership, family, associates, employment, court-case parties, sanctions) carry start and end dates and a summary that states the link and its strength on the yardstick, for example “director of (Finance Director): likely (55–75%)”. Phones, emails and domains are attached to the party they belong to. Exporting again updates rather than duplicates. What FollowTheMoney has no place for (IP addresses, certificates, open ports) is counted in the result and not written.

From them. Import reads such a file, including OpenSanctions and Aleph exports. The file is stored as evidence (manually retrieved, by you) and every entity and link cites it, so the usual provenance and export rules apply to imported material. People, organisations, addresses, properties, vehicles, articles, court cases and events become entities, and the link kinds above become relations. Any other kind of record (vessels, aircraft, securities, passports and so on) is skipped and counted by name in the result, as are lines that are not FollowTheMoney and links whose end was not imported. A Viokit export imported back restores identifiers, roles, qualifiers and dates. An imported foreign file has no Viokit ratings: its links are unrated, not certain.

To Maltego. Export writes a GraphML file (in Maltego: Import Graph from GraphML) and a CSV with one row per link: source type and value, target type and value, link label. People, organisations, addresses, phones, emails, domains, IPs, URLs, articles and usernames use Maltego's standard entity types. Properties, vehicles, events, court cases and the rest are custom types, which Maltego shows generically until you define them. Labels read like “director of (Finance Director), likely (55–75%)”.

What stays behind. Both exports honour the evidentiary export's withholding. A step citing an artifact that has been redacted, or that was read from a restricted or sensitive offline dataset you did not choose to include, is left out together with the entities and links it asserted, and the result says how many steps and which artifacts. Check what was withheld before sharing.

Births, marriages and deaths

Country Record Access Viokit
England & Wales GRO birth/death/marriage indexes (1837–) GRO online index (free account, no API); certificates paid manual
England & Wales FreeBMD (volunteer transcription of the GRO index, to about the 1990s) free search, no API manual
England & Wales Deaths: The Gazette deceased-estates notices (section 2900s) open feed built (notices about companies; no live query on a named person)
England & Wales Probate (will and grant search on GOV.UK) free search, no API, documents paid manual
Scotland ScotlandsPeople paid, no API manual
United States State vital records; SSA Death Master File (restricted since 2014) varies; DMF restricted no
Notable people Wikidata birth/death dates and places, spouses, parents, children open SPARQL built
Russia ZAGS civil registry not public no

Criminal records and courts

There is no lawful public criminal-records database in the UK: DBS checks need the subject's consent, and the Police National Computer is closed. Criminal matters surface through:

Source What it gives Viokit
Find Case Law (The National Archives) High Court, Court of Appeal and tribunal judgments, searchable built
Court lists (Crown/magistrates, CourtServe) who is listed for hearing, by court and date manual (no API; terms restrict scraping)
News coverage (GDELT, The Guardian) reporting of arrests, charges, convictions built
Insolvency Service: disqualified directors Companies House API disqualified-officers next (on the Companies House pack)
US: CourtListener / RECAP (Free Law Project) federal court dockets and opinions (free API, token) built (needs a key)
US: PACER federal dockets (paid per page) no
Russia: sudrf.ru, kad.arbitr.ru (commercial courts) court decisions and arbitration cases catalogued

Spent convictions under the Rehabilitation of Offenders Act 1974 should not be relied on in employment or similar contexts even when found in old news; analysts should note the date of anything found.

Property, land and sales

Country Record Access Viokit
England & Wales HM Land Registry Price Paid Data (every residential sale since 1995) open, SPARQL + bulk CSV (OGL) built
England & Wales UK companies that own property (CCOD) and overseas companies (OCOD) free with account, bulk CSV next (DCAT manifest; download needs the owner's account)
England & Wales Title register and plan for one title £7 per document via GOV.UK manual
England & Wales INSPIRE index polygons (freehold title extents) open bulk by local authority next
England & Wales Planning applications per local authority portals manual
Scotland Registers of Scotland paid manual
United States County assessor and recorder (deeds, mortgages, owners) varies by county; many free search sites catalogued
Russia Rosreestr (EGRN extracts) paid extracts catalogued

How it joins up: a sale is linked to its address, keyed the same way as a Companies House registered office, so one address shows both the companies registered there and the sales of it.

Vehicles

Source What it gives Viokit
DVLA Vehicle Enquiry Service API (free key) make, colour, fuel, first registration, tax and MOT status for a registration mark built (needs a key)
DVSA MOT history API (free key) MOT tests, mileage, defects built (needs a key)
Keeper details DVLA, "reasonable cause" only no

News, publications and reputation

Source Access Viokit
GDELT DOC 2.0 (worldwide news index, 65 languages) open built
The Guardian Open Platform free key built (needs a key)
Common Crawl News open bulk (large) no (disk)

Sanctions, PEPs and watchlists

Source Access Viokit
UK Sanctions List (FCDO) open data pack (screening)
OFAC SDN (US Treasury) open data pack
OpenSanctions (FtM, 300+ lists incl. PEPs) bulk free for non-commercial use data pack (PEPs)
EU Financial Sanctions File open (account) data pack

Lines of investigation (playbooks)

Viokit can run a named line of investigation (a playbook; see Playbooks) rather than a raw tree:

  1. Who is behind this website? domain → infrastructure, page contacts → registered company → officers → their other appointments → shared offices.
  2. Corporate due diligence. company → registry profile, officers, PSCs, filings, parents (GLEIF, Wikidata), insolvency notices, judgments, press, sanctions screening.
  3. What has happened at this address? postal address → companies registered there, sales and prices, insolvency notices naming it.
  4. Domain infrastructure. domain → DNS, subdomains, certificates, hosting network and owner, open ports (scanners only against authorised targets).
  5. Person footprint (opt-in). a named person → registry appointments, judgments, press, Wikidata facts, username accounts. Needs the owner's go-ahead for each private individual.

Coverage by jurisdiction

Records are national. The tree places each entity in a jurisdiction (registry keys, postcodes, phone calling codes, ccTLDs), skips sources that do not cover it, and reports in its summary which record types went dark and the manual source to use instead.

What is deliberately not built

  • Leaked or breached personal data, bought or found. A privately held dataset must meet strict safeguards before it is connected.
  • The open electoral register (recommended no: the owner has not decided).
  • Paid people-search aggregators.
  • Scraping services whose terms forbid it (LinkedIn, court lists).
Docs

↑↓ to moveEnter to openEsc to close