v1Five-safeguard screening engine

Sanctions, PEP & adverse-media screening you can prove

Screening across 1.74M sanctioned, PEP and enforcement records — including1,031,563 politically-exposed persons — every decision sealed into WORM cryptographic evidence with RFC 6962 audit proofs. Screen against a non-overridable sanctions floor, then show an examiner exactly what was checked, when, and why.

1.74M
watchlist records
1,031,563
politically-exposed persons
216
official sources
RFC 6962
audit proofs
OFAC · UN · EU · UK HMT consolidated listsPEP & RCA coverageAdverse-mediaData-residency awareAppend-only audit trail1.74M entities · full coverage →
The engine

Five safeguards between a name and a decision

Screening you can defend is not one model — it is layered controls that each fail closed. Every one of them is enforced server-side and written to the evidence chain.

Non-overridable sanctions floor

Hard-sanctions matches surface at the platform floor regardless of any sensitivity slider. No UI control, key, or role can lower it — the floor is enforced in the engine, not the interface.

Independent best-of-three checker

Each result is re-derived by an independent scorer and reconciled across three passes. Disagreements are flagged, never silently dropped.

Maker-checker adjudication

Releasing or dismissing an asset-freeze match requires a distinct second approver. You cannot approve your own disposition — the platform refuses it.

Hash-chained evidence capsules

Every screening seals the full candidate set, scores and list snapshot into a capsule, hash-chained to the last and archived to WORM storage.

Metered pay-per-screen

Billing is counted from the same immutable decision log you are audited against. Start free with 200 screens a day; beyond that, subscribe to a plan or pay $1.50 per screening — no contract required either way.

WAF- & DDoS-hardened API

A rate-limited, WAF-fronted REST API with scoped ssk_ keys, signed webhooks and per-tenant isolation. Test and live keys never mix.

Developers

One call. A decision you can hand to an examiner.

A single REST endpoint returns a decision, the scored candidate set with a factor breakdown, and the id of the sealed evidence capsule.

  • Scoped keys. Mint ssk_test_ and ssk_live_ keys from the console, each limited to the scopes it needs.
  • Structured results. Every match carries its raw and adjusted score plus the exact factors that moved it.
  • Provable by id. The audit_id resolves to a WORM-sealed capsule with the full decision fingerprint.
screen_entity.sh
curl -X POST https://api.celmus.com/v1/screen/entity \
  -H "Authorization: Bearer ssk_live_…" \
  -H "Content-Type: application/json" \
  -d '{
    "subject": {
      "entity_type": "individual",
      "name": "Alexander Volkov",
      "dob": "1984-03-12",
      "nationalities": ["CY"]
    }
  }'

# → {
#     "decision": "review",
#     "matches": [{ "canonical_name": "…",
#        "score_adjusted": 94, "restriction": "…" }],
#     "audit_id": "a7f3c9e2-…",
#     "capsule": { "state": "sealed" } }
Who screens with celmus

Built for anyone who has to prove they checked

Sanctions exposure is not only a banking problem. If a regulator, an auditor or a counterparty can ask who did you screen, against what, and when, celmus is built to answer it. The obligation stays yours — celmus is the infrastructure and the evidence behind the answer.

Financial institutions

Customer due diligence at onboarding and re-screening thereafter, with structured PEP occupancy and relatives/close-associate links so enhanced due diligence is evidenced rather than assumed. Banks, lenders, credit unions and money services alike — screen sender and beneficiary before funds move, for FINTRAC, FinCEN and equivalent regimes.

Payment service providers

One REST call per entity, or batch your whole book. Sandbox keys in minutes, the same corpus as production, and per-tenant isolation if you screen on behalf of your own merchants — their decisions stay in their own evidence trail, not yours.

Insurance companies

Screen the policyholder at underwriting and the beneficiary at claim — including the beneficiary who was changed after the policy was written, which is the event that actually carries the risk. Corporate policyholders resolve through the ownership graph to the people behind them.

Crypto businesses

Sanctioned wallet addresses are first-class entities alongside people and companies, so a VASP screens the counterparty and the destination address through one control and gets one decision record covering both.

Gaming & gambling

Player onboarding has to clear in seconds without lowering the bar. A sub-second screen returns a decision at signup, PEP and adverse-media hits route to review instead of blocking silently, and every deposit threshold you must evidence has a capsule behind it.

Real estate agencies

Buyer, seller, and the beneficial owners of whichever company is really purchasing. The ownership graph carries control edges as well as shareholding, so a nominee structure resolves to the person who ultimately benefits.

Asset management

Investor onboarding and periodic re-screening of the whole book, with the UBOs of funds, trusts and nominee holdings resolved rather than accepted at face value. Batch the book on a schedule; only what changed needs an analyst.

Business intelligence

Bulk entity resolution over sanctions, PEP, debarment and adverse-media sources with full provenance — every match carries the source, its licence basis and the date it was published, so research output can be cited rather than merely asserted.

Procurement and third-party risk are covered by the same corpus — US federal exclusions and multilateral development-bank ineligibility sit alongside sanctions, so supplier onboarding catches barred vendors and not only sanctioned ones. Vessels and aircraft are first-class entities too, for trade finance and shipping exposure. Not sure whether your obligation fits? Start free — 200 screens a day, no card — or ask sales@celmus.com.

Pricing

Monthly subscriptions. PEP included. No seats.

Start free with 200 screens per day — no card. Past that it is $1.50 per screening, or a plan below with an included monthly allowance at a much lower effective rate. Allowances reset each month and do not roll over. Every tier includes sanctions, PEP, enforcement and adverse-media.

Free
Evaluate and build. No card, no contract.
$0
200 screens / day
  • Full engine — sanctions, PEP & enforcement
  • Same corpus as paid tiers
  • Test keys & evidence capsules
  • Then $1.50 / screening, or subscribe
Starter
Go live. Live keys, entity + batch + transaction, PEP included.
$799/ month
2,200 scans / month · non-rolling
  • Sanctions, PEP & enforcement screening
  • Entity, batch & transaction
  • WORM evidence capsules & audit export
  • Maker-checker case workbench
Growth
More volume + adverse-media.
$1,299/ month
6,500 scans / month · non-rolling
  • Everything in Starter
  • Adverse-media screening
  • Higher rate limits & batch size
Scale
High volume + continuous monitoring.
$1,999/ month
15,000 scans / month · non-rolling
  • Everything in Growth
  • Continuous monitoring & re-screening
  • Batch up to 2,500 / call
Enterprise
Regulated volume, residency & procurement.
Custom
committed volume · data residency
  • Region-pinned data residency
  • SSO, custom RBAC & audit-lead roles
  • SLA & onboarding support

Screening is easy. Proving it is the product.

Every decision celmus makes is sealed, hash-chained and independently verifiable — so when an examiner asks what you checked, the answer is already written down.