Skip to main content
The search history endpoint returns the screening requests your tenant has run. Use it for audit retrieval: a compliance team can reconcile which screenings ran, when they ran, who ran them where Minerva recorded a user, and which submitted entities and feeds each request included.
Search history retrieval is read-only. It does not submit a new screening, change review status, or modify historical records.

Endpoint

Authenticate with the same API key header the other Minerva API endpoints use:
The API key determines the tenant whose historical requests are returned. Create and manage keys from Administration > Developers in the Minerva dashboard, or see API Keys.

Query parameters

Use RFC3339 date-time values with an explicit timezone, such as 2026-05-01T00:00:00Z. When you build URLs, use your HTTP client’s query parameter encoder, URLSearchParams, or --data-urlencode, so reserved characters such as : and + are encoded safely.

List recent search requests

Retrieve a date range

Date range filters are inclusive and apply to the request’s created_at timestamp.

Example response

Paginate an audit export

For a full audit export, request pages until pagination.has_next is false. Keep the same filters on each page so the result set remains consistent.
JavaScript

What changed since the last run

Monitoring screens each monitored profile again on its cadence. To see what the latest screening found, use the searchId from the profile_status_updated webhook and list that screening’s unresolved matches:
cURL
This works because of repeated alert suppression. With suppression on, a match that repeats one from the profile’s recent screenings on the same feed (the last 3 by default, set by Previous searches to scan) is marked suppressed instead of unresolved, and a screening where every match is suppressed sends no webhook. So review_status=unresolved on the latest screening returns the matches that suppression did not treat as repeats. These are usually new findings, plus any repeat that didn’t meet the suppression rules. Treat it as a close approximation, not an exact difference. Suppression is on by default with the Balanced preset, and you configure it per workspace in the Minerva dashboard. If it is off in a workspace, every match in each screening is unresolved, including repeats. You can also query by profile and time window:
Minerva cannot list matches that disappeared since the previous screening. To find them, compare the matches of two screenings in your own system.

Monitoring run history

To show that a monitored profile was screened, including screenings that found nothing, list its search requests:
cURL
Each item in requests[] is one screening request for the profile. A request is recorded for every screening, whether or not it found anything.
  • Monitoring screenings carry config.billable_tag set to monitoring. Other screenings of the profile, such as the onboarding screening, appear in the same list with a different tag or none.
  • created_at is when the screening was requested.
  • entity_results is keyed by the index of each submitted entity, starting at "0". An empty object, such as {"0": {}}, means every feed answered. When a feed did not return a complete answer, the entity’s entry carries feed_statuses.<Feed>.status set to partial or unavailable, and partial: true. That feed was not fully screened in that run.
The request does not carry a result count. To get one, ask for a single match from that request and read the total:
cURL
pagination.total_count is the number of matches that screening produced. 0 means it found nothing.
A request has no separate completion time. Use created_at as the time of the screening.

Retrieve associated results

The search history response gives you the request metadata and the request identifier. When you need the stored match-level results for a historical request, use the returned id as the request_id on the search matches endpoint.
cURL
The response returns the stored potential match objects. They use the same field layout as a direct search response, so the field descriptions and the recommended analyst review order in the Screening Integration Guide apply to them as well. The endpoint reference lists every filter and the pagination fields.

Match IDs

Each stored match carries an id. It is the value you send as match_id to update the match’s review status, so you can save it in your system and record a decision later. A synchronous POST /v2/search returns the same value on each match as screening_match_id, so you can take the ID straight from the response without reading the matches back. If a match in the response has no screening_match_id, read the request’s matches with GET /v2/search/matches?request_id=<request_id> and use their id values.
  • Stable. A match keeps its id for as long as it exists. Updating its review status doesn’t change it.
  • New per screening. A new screening, such as a monitoring rescreen, stores its matches with new IDs and its own request_id. Minerva never reuses the ID of an earlier match, so if a rescreen finds the same person or entity again, that potential match has a new id. On monitored profiles, repeated alert suppression can mark a match that repeats an earlier finding as suppressed.
  • Retained. Minerva doesn’t expire stored matches automatically, and archiving a profile doesn’t remove them.
  • Not the record ID. A match can also carry record_id, the data source’s own identifier for the matched record. It isn’t a match ID, so don’t send it as match_id.
When a synchronous POST /v2/search returns an entity in entity_matches, the matches listed for it are already stored. You can read them with GET /v2/search/matches and the request_id from the same response. In rare cases Minerva completes the search but can’t store an entity’s matches. That entity’s metadata then includes results_stored: false. The field is absent on normal responses and doesn’t change the status code or any other field. When you see it, that entity’s matches may be missing from GET /v2/search/matches. If you need them stored, run the search again. An asynchronous search returns a job_id and no entity_matches, which means the search is still running. Read its matches with GET /v2/search/matches and the request_id once the job has completed.