What does Verifex screen against?

Verifex screens people and companies against sanctions and watchlists, including OFAC, UK Sanctions, the United Nations, and BIS. Coverage and update frequency vary by source; see the data sources page for details.

What does it mean that a decision can be replayed?

Every screening records the data, matching evidence, source snapshot, and engine version behind the result, tied to a decision hash. That lets your team reconstruct and verify what was recorded — the versions, decision identity, and source provenance of a past decision — instead of relying on a screenshot or the final label alone. (Metadata and provenance replay; exact historical engine re-execution is not currently available.)

What do the benchmark numbers mean?

The figures come from Verifex's published 500-case synthetic benchmark (engine v3.0.0, verified May 17, 2026): 99.68% F1, 99.36% recall, 100% precision, with 0 false positives and 3 false negatives. This is a benchmark result on synthetic cases, not a guarantee of real-world performance.

Screening assurance infrastructure

Screening assurance infrastructure to test, replay, explain, and defend sanctions and PEP decisions.

Test, replay, explain, and defend compliance screening decisions with the data, engine version, matching evidence, and decision context behind every result.

Delivered through one developer-first API.

Explore a decision record Start with the API

Developer-first API · Reproducible evidence· Versioned decisions

One decision, five surfaces

See exactly what stands behind a screening result.

The same record can be opened as a decision, its evidence, a replayable identity, a version comparison, and an audit trail — the difference between a label and a defensible decision.

Decision Assurance Console

Synthetic assurance walkthrough

Screening query

Northwind Maritime LLC · entity

risk_level: high

candidate: NORTHWIND MARITIME LLC
source: OFAC SDN
match_type: FUZZY
signals.match_strength: strong
confidence: 94
adjudication.action: ESCALATE
lists_checked: OFAC · UK · UN · BIS

A screening result is not enough.

A match may be reviewed months later, after sanctions lists, aliases, policies, and matching engines have all changed. By then a final label answers nothing. Teams need to reconstruct and verify what was recorded — the data, evidence, versions, and source provenance behind the decision — not merely retain the outcome.

Core assurance workflow

Test, replay, explain, defend — one connected system.

Each stage feeds the next. Together they turn a single screening call into a decision your team can stand behind.

01
TEST
Run controlled cases and verify how the matching engine behaves before it reaches production.
02
REPLAY
Verify the recorded versions, decision identity, and source provenance behind a past result. Metadata and provenance replay; exact historical re-execution is not currently available.
03
EXPLAIN
Inspect the sources, matched fields, confidence, and evidence behind why a result was produced.
04
DEFEND
Preserve a reviewable decision record that holds up to internal and external scrutiny.

Evidence-backed decision record

Every screening leaves a record you can open later.

One compact decision record captures the query, the candidate, the list and source snapshot, the matching signals, the engine version, and the recommendation — retrievable by decision ID so a result can be inspected long after it was made.

How decision records work

Decision recordSynthetic

query: Northwind Maritime LLC
candidate: NORTHWIND MARITIME LLC
source: OFAC SDN
matching signals: name match · confidence 0.94
engine version: verifex-screening-v5.0
source snapshot: ofac_sdn · 6b1f…c904
recommendation: escalate for review
decision_hash: sha256:2d7e…a41f

Developer integration

One request. A structured, evidence-backed response.

Screen a party through a documented REST API and get back the matches, the sources checked, and a decision hash you can retrieve later — from your own stack.

Request

POST https://api.verifex.dev/v1/screen
Authorization: Bearer $VERIFEX_API_KEY
Content-Type: application/json

{ "name": "Northwind Maritime LLC", "type": "entity" }

ResponseIllustrative example

{
  "request_id": "9f2a4c7e1b3d5a8c",
  "risk_level": "high",
  "total_matches": 1,
  "matches": [
    {
      "primary_name": "NORTHWIND MARITIME LLC",
      "source": "OFAC",
      "entity_type": "entity",
      "match_type": "strong",
      "confidence": 94,
      "risk_level": "high"
    }
  ],
  "lists_checked": ["OFAC", "UK", "UN", "BIS"],
  "api_version": "v1",
  "decision_hash": "sha256:2d7e…a41f"
}

Read the docs

Benchmark & verification

The “test” half of the workflow, measured in the open.

Matching quality is checked against a published fixture set before a change ships, so you can see how the engine performs rather than take it on trust.

Benchmark result on synthetic cases, not a guarantee of real-world performance. Adversarial recall varies by corruption type, and we do not publish latency percentiles until an independently verified load test is reconciled.

Benchmark methodology

99.68%

99.36%

Recall

100%

Precision

0 / 3

False positives / negatives

Sample: 500-case published benchmark
Engine: v3.0.0
Verified: May 17, 2026

Sources

OFAC
UK Sanctions
United Nations
BIS

Coverage and update frequency vary by source.

Security posture

Encryption in transit, hashed API keys, audit logging, and a documented vulnerability-disclosure process. See the security and legal pages for current controls and limitations.

Security Privacy DPA Subprocessors

Make every screening decision reproducible.

Get an API key Talk to Verifex

Free developer plan · documented API and SDKs· evidence on every decision.