award_counterparty_state

Product · v0 · shipped 2026-09-05

award_counterparty_state

A company won a real EU public contract — is it still alive today? Given a TED award notice, this resolves the winning company's real commercial-registry status right now: active, dissolved, or in insolvency proceedings. Same engine as find_missing, one new source added (Spain's BORME history via OpenMercantil) — nothing hidden, nothing invented.

A real bug was found and fixed (2026-09-07 → 2026-09-10)

A precision test run on 2026-09-07 (comparing this product's output against independently-verified ground truth, not just the source it already queries) found 3 real, confirmed cases — out of 38 real EU award notices checked — where the product reported a confident "confirmed" active status while attributing it to the wrong company (a syntactically valid but incorrect identifier resolved to an unrelated business, e.g. a French award to "VALLOIS SAS" returned the registry data of an unrelated Paris real-estate micro-company). That is worse than an honest "not found" — it produces false confidence. Paused across all 4 countries (ES, FR, GB, HU) the same day this was found, including Spain, where no bad case happened to land in this sample — the underlying bug (no resolver checked the returned company name against the real award winner) was structural, not country-specific.

The root-cause fix (comparing the registry's returned name against the real TED winner name before reporting a confident status, instead of trusting a syntactically valid identifier alone) was written, tested, and deployed the same day — re-running the exact same 38 real cases plus 15 known-active control companies came back clean in all 4 countries: 0 identity errors, 17/17 new tests, 107/107 full suite. The product stayed paused regardless until an explicit re-activation decision — re-activated 2026-09-10.

Try it — real response, right now

curl https://api.nkodatalabs.com/v0/award_state/15-2026

Real response for a real TED award (a Spanish municipal contract) — registry_status.active: true and registry_status.name_matches_ted_winner: true (the exact check the 2026-09-07 fix added), sourced from BORME via OpenMercantil, with a link to the evidence:

{
  "winner": {"name": "ELECNOR SERVICIOS Y PROYECTOS SAU", "country": "ESP", "identifier": "A79486833"},
  "registry_status": {
    "source": "OpenMercantil (BORME)",
    "status": "resolved",
    "real_name": "ELECNOR SERVICIOS Y PROYECTOS SA",
    "registry_status": "Activa",
    "active": true,
    "name_matches_ted_winner": true,
    "evidence_url": "https://openmercantil.es/empresa/elecnor-servicios-y-proyectos-sa"
  }
}

Real coverage, published honestly

Measured live on a real 25-award sample per country, last 12 months (2025-09 to 2026-09), chosen by TED's own API — not cherry-picked. Full breakdown and methodology in the repo (award_counterparty_state_resolution_rate.md).

CountryResolution rateSource
Spain (ES)65% (13/20 real companies, 5 natural persons correctly excluded)OpenMercantil / BORME
France (FR)72% (18/25)SIRENE
UK (GB)32% (7/22)Companies House
Hungary (HU)0% — no real free public registry API existsnone available today

Hard rule: this never asserts fraud, illegality, or bid-rigging — only the winning company's real registry status today, with the exact source and evidence link.

Get a free key (50 calls/day)

Self-service, no approval — POST /v0/keys with your email returns a key instantly. Shared quota with find_missing and /v0/company_state.

MCP tool

Real MCP server (streamable-http), tool name award_state, same server as find_missing: https://api.nkodatalabs.com/mcp. Works keyless within the same anonymous trial.

Pricing

Same shared pricing as find_missing: free (5/day no key, 50/day with a key), then 2,500 calls for €19, never expires.