Does SanctionsAI support multi-language screening?

SanctionsAI screens SDN entries across multiple scripts, including Latin, Cyrillic, Arabic, and Chinese, with fuzzy matching for transliteration.

TL;DR

TL;DR: Yes. SanctionsAI screens names in multiple scripts and uses fuzzy matching to catch transliteration variations, so a name spelled differently still matches.

Why scripts matter

The SDN List contains names written in several scripts. A name in Cyrillic or Arabic may transliterate into Latin in many ways, and a naive exact match will miss most of them. Screening that handles the original script plus transliteration reduces false negatives, which is where sanctions evaders hide.

How SanctionsAI handles it

SanctionsAI screens against 19,218 SDN names using fuzzy matching that accounts for transliteration variations across Latin, Cyrillic, Arabic, and Chinese scripts. The check returns in under 100 ms, so multi-language matching does not slow the payment path.

What this means for your agents

An agent screening a counterparty gets a clean ALLOW or BLOCK result regardless of the script. The 4-Gate protocol still applies: SCREEN, SCORE, STOP, STAMP. Log every screen, and block on any match, including matches found through transliteration.

Transliteration is where matches hide

The same name can appear as many different Latin spellings, and an exact match misses all of them. Fuzzy matching that understands how Cyrillic and Arabic names transliterate closes that gap. For a screening tool, coverage of scripts and transliteration is not a nice-to-have; it is the difference between catching a designated name and letting it through. For a global product, that coverage is the difference between a robust screen and one that only catches names written in Latin.