Early accessNow onboarding design partners.Request access
AutoCompliance

Sanctions screening API

Sanctions screening that understands names

Catch transliterations, typos, reordered and partial names, without burying your analysts in false positives. Two-stage AI matching, explainable scores and one API for every list.

  • EU company, GDPR by design
  • Optional zero data retention
  • Sandbox with test lists
POST /v1/screenings
curl https://api.auto-compliance.com/v1/screenings \
  -H "Authorization: Bearer ak_test_..." \
  -H "Content-Type: application/json" \
  -d '{
    "full_name": "Mohamad al-Rashid",
    "date_of_birth": "1971",
    "external_id": "cust_83921"
  }'
200 OKhit
{
  "hit": true,
  "threshold": 75,
  "matches": [{
    "score": 94,
    "matched_name": "AL RASHID, Mohammed",
    "score_breakdown": {
      "retrieval": 0.88,
      "verifier": 0.97,
      "date_of_birth": "partial"
    },
    "entity": {
      "source": "ofac_sdn",
      "list_version": "2026-11-19"
    }
  }]
}

Illustrative sandbox response.

Official sanctions lists, refreshed daily and versioned

The problem

Exact matching misses people. Naive fuzzy matching floods your queue.

The same person is written many ways across passports, payment messages and sanctions lists. Rule-based variations grow exponentially, and loose string similarity flags everyone with a common name.

Transliteration

Mohammed Al Rashid

Muhamad al-Rasheed

Same person

Reordered names

Ivanov Ivan Petrovich

Ivan Ivanov

Same person

Diminutives

Alexander Volkov

Sasha Volkov

Same person

Typos & titles

Dr. Jonathon Smyth

Jonathan Smith

Same person

And the reverse: Maria and Marina, or Alexander and Alexey, look alike as strings but are different names.

How it works

Two stages: wide recall, then precise verification

  1. 01

    Normalize

    Titles, punctuation and casing are removed. Name parts are tokenized, so order does not matter.

  2. 02

    Retrieve

    Character n-gram vectors find phonetically and orthographically similar candidates across millions of names in milliseconds.

  3. 03

    Verify

    An AI language model compares each candidate with the input and filters out look-alikes that are different names.

  4. 04

    Explain

    The final score combines both stages with date of birth. You see why every match was returned.

Learn more about the matching engine →

Platform

Everything a compliance team needs to screen people

Real-time API

Screen a person in a single synchronous call, built for onboarding and payment flows.

Batch screening

Upload up to 100,000 records as CSV, Excel or JSON. Poll the status or get a webhook when the batch is done.

Explainable scores

Every match shows the retrieval score, the AI verifier score, the date-of-birth signal and the alias that matched.

Audit-ready history

Each result records the exact list version and entry version, so you can reproduce any decision later.

Coming soon

Custom lists & whitelist

Screen against your internal blocklists with the same algorithm. Suppress confirmed false positives until the entry changes.

Coming soon

Bring your own provider

Already licensed with a data vendor? Index their bulk data alongside official lists at a reduced rate.

Security & privacy

Built for regulated businesses

AutoCompliance is operated by a Cyprus-registered EU company. Personal data stays in the EU region, is encrypted in transit and at rest, and is never used to train models without your consent.

Visit the trust center
  • GDPR by design

    Configurable retention and deletion on request.

  • Zero-retention mode

    Opt out of storing inputs and results entirely.

  • Roles & 2FA

    Admin, analyst and viewer roles with mandatory 2FA.

  • Full audit trail

    Every decision is logged with who, when and why.

Screen smarter from day one

We are onboarding a small group of design partners. Tell us about your screening volume and lists, and we will set up sandbox access.