Skip to main content
Intelligence

data.ndocs: what the same person has done with you, in every answer.

Markdown

Every finished process tells you who the person is, with the Human Token, and what that person has done with you, with the intelligence in data.ndocs: with how many of your clients they were seen, how many documents they used, and when.

"ndocs": {
"distinctClientsCount": 2,
"transactionsCount": 3,
"firstSeen": "2026-09-08",
"lastSeen": "2026-11-03",
"credentials": [
{ "type": "DUI_TYPE_NG_BVN", "countryCode": "NGA", "distinctCredentialCount": 1 }
]
}
FieldMeaning
distinctClientsCountDistinct clientId values across the person's processes
transactionsCountNumber of the person's processes, this one included
firstSeen, lastSeenUTC dates (YYYY-MM-DD) of the earliest and the latest process
credentials[].typeDocument type, a code of the attribute catalog
credentials[].countryCodeAlpha-3 country that issues the document. Absent for a type the catalog no longer lists
credentials[].distinctCredentialCountDistinct documents of that type

Reading it​

A distinctClientsCount above 1, or a distinctCredentialCount above 1, means the same face has been used with different clients or documents. What to do with that signal is your decision.

What is counted​

A person is identified by the Human Token. Only finished processes with that token are counted, across all your client ids. A process created without a document counts as a transaction and as a client, but under no document type. credentials is [] when none of the person's processes has a document.

The intelligence covers the processes of your Human Plugin installation.

Whose intelligence it is​

  • Identification that finished with a Human Token: that person. Not when data.liveness.value is not_live.
  • Revalidation with data.humanTokenMatch equal to "true": the person of the token you sent in references. The process counts for that person.
  • Otherwise (a negative match, or a capture that is not live): no intelligence, and no data.ndocs.

When it is absent​

data.ndocs can be missing for a known person:

  • the process is not finished yet;
  • the intelligence could not be computed. The create or read still succeeds, and a later GET tries again.

Do not read a missing data.ndocs as a person never seen.

Next​

Prerequisites and access