Agentic KYB — EU Company Verification (VAT + LEI + Sanctions)

One call answers: can my agent (or my business) safely transact with this company?

Give it a VAT number, an LEI, or a company name — get back a single structured record combining:

• ✅ EU VAT validation via VIES (the European Commission's official system) — validity plus the registered name and address where the member state discloses them
• 🏛️ Legal entity verification via GLEIF — the global LEI record: legal name, jurisdiction, entity and registration status
• 🚫 Sanctions screening against the EU consolidated financial sanctions list and the US OFAC SDN list — refreshed daily, with transparent match confidence
• 🚦 A composite trust flag: CLEAR / CAUTION / REVIEW / HIT / INSUFFICIENT_DATA, with human-readable reasons

Built for AI agents (callable as an MCP tool, flat JSON output, deterministic behavior) and equally useful for invoicing automation, B2B onboarding, and e-commerce checkout checks.

Pricing

$0.05 per company check (pay-per-event). A check that fails entirely — every source unavailable — is not charged. Typical commercial KYB checks run $0.20–$3.00; enterprise platforms start at $500/month. Scheduled re-checks in monitoring mode bill on a separate, cheaper monitoring-check event — see below.

Why this instead of a plain VAT validator?

A VAT check tells you a number is valid. It does not tell you the counterparty is safe. This actor answers both in one call, for a fraction of the cost of traditional KYB platforms — typical commercial checks run $0.20–$3.00 per screen; enterprise platforms start at $500/month.

Input

{
  "companies": [
    { "vatNumber": "IE6388047V" },
    { "name": "Sberbank of Russia" },
    { "lei": "HWUPKR0MPOU8FGXBT394" }
  ]
}

Each item may combine vatNumber (with country prefix), lei, and name (+ country). More identifiers = stronger verification. Single-check convenience fields (vatNumber, companyName, lei, country) also work.

Output (one record per company)

{
  "trustFlag": "CLEAR",
  "trustReasons": ["VAT number valid in VIES", "No sanctions matches (OFAC_SDN, EU_FSF)"],
  "vat": { "status": "ok", "valid": true, "name": "GOOGLE IRELAND LIMITED", "address": "…" },
  "lei": { "status": "found", "lei": "…", "legalName": "…", "registrationStatus": "ISSUED" },
  "sanctions": { "listsChecked": ["OFAC_SDN", "EU_FSF"], "matches": [] },
  "checkedAt": "2026-07-29T21:04:00.000Z"
}

Sanctions matches always include the list entry, the name matched against, and a confidence of high (exact normalized match) or medium (token-based — review advised).

EU sanctions list token (free, optional)

OFAC screening works out of the box. To include the EU consolidated list, supply a free EU FSF download token (create an EU Login, request file access at the Commission's Financial Sanctions Files site) via the euFsfToken input field. Without it, screening transparently reports OFAC-only coverage.

Honest limitations

• Name matching is name matching. It produces false positives (common names) and false negatives. medium matches are flags for review, not verdicts. Cyrillic and Greek names are transliterated (BGN/PCGN-style) so they screen against Latin queries and vice versa; Arabic and CJK scripts are not transliterated — no deterministic letter map exists, and a wrong guess is worse than an honest gap. Transliteration maps script, not language: "Россия" matches "Rossiya", not "Russia".
• VIES member-state backends throttle and occasionally go down; the actor retries and reports unavailable honestly per country rather than guessing. Germany and Spain do not disclose names/addresses via VIES.
• Query-time screening only by default — see Monitoring mode below for scheduled re-checks.

Monitoring mode

Set monitoringMode: true and re-run the same companies on a schedule (e.g. Apify Scheduler, daily or weekly) to get change alerts instead of re-reading full records yourself. Each result gets a monitoring field:

"monitoring": {
  "key": "vat-IE6388047V",
  "changed": true,
  "isFirstCheck": false,
  "changes": ["trust flag changed: CLEAR → HIT", "new sanctions match(es): EU_FSF:12345"]
}

The first run for a given company is isFirstCheck: true (nothing to compare against yet — it establishes the baseline). Subsequent runs compare trust flag, VAT validity, and sanctions matches against the last check. Billed on a separate, cheaper monitoring-check event rather than company-check.

Disclaimers

Unofficial service. Not affiliated with or endorsed by the European Commission, GLEIF, or the U.S. Treasury/OFAC. Results are a screening aid from official public sources — not legal advice and not a substitute for regulated AML/KYB obligations. Data minimization: results are returned to you and not retained beyond the run's dataset; sanctions lists are cached for 24 h. Registry data on sole traders may constitute personal data — process it under your own lawful basis.