1. Official source data
The backend downloads structured sanctions data directly from four authority endpoints: the European Commission's consolidated financial sanctions data, the OFAC SDN List, the UK Sanctions List and SECO's overall Swiss sanctions data. Source-specific parsers convert the different XML and CSV formats into a common internal representation while retaining the source name, official reference, dates, names, aliases and available sanctions descriptions.
The service attempts a full refresh when it starts and every hour afterwards. A legal decision or measure may take effect before the responsible authority updates its searchable or downloadable dataset, and the service can retrieve a change only after the authority publishes it digitally. A failed connection, altered data format or parsing error may create an additional delay. The refresh schedule is therefore not a guarantee of completeness or real-time information.
2. People and legal entities
Records are classified as people or legal entities from the subject type supplied by the source. Searches use the selected category to avoid comparing a person query with every organization record or vice versa. Authorities can publish other categories, such as ships and aircraft, that are not separate searchable categories in the current app.
3. Names and aliases
The matching process considers the main official name and the alternative names parsed from each source. For comparison, text is converted to a common uppercase form, diacritic marks are removed, punctuation becomes a separator and repeated words are ignored. A limited set of common corporate suffixes, such as LTD, LLC and GMBH, is excluded from token comparison so that legal-form abbreviations do not dominate an organization score.
For Cyrillic input, the matcher also creates a bounded set of Latin transliteration variants. These are technical search aids, not official aliases. An official spelling supplied by the authority is preferred when scores are otherwise equal.
4. Similarity calculation
Names are compared using edit-distance similarity and token-aware comparison. Multiword searches assign each query word to a different candidate word, which helps with reordered names while preventing the same candidate word from satisfying several query words. Longer query words carry proportionally more weight than shorter ones.
The resulting value is rounded to a percentage and filtered by the selected threshold. The web screening endpoint currently uses a threshold of 80%. Results are sorted by similarity and limited to 100 records.
5. What the score does not include
The percentage is based on the name text. It does not incorporate dates of birth, nationality, addresses, ownership, passports, company registration numbers or the legal reasons for designation. It is therefore not:
- a probability that the searched subject is sanctioned;
- a legal or compliance risk score;
- a measure of sanctions severity;
- proof that two records identify the same subject.
6. Online and offline processing
When an online search is used, the entered phrase is sent to the Sanctions Sleuth backend so the service can calculate results. The Android app may also use downloaded offline data for supported offline behaviour. The current Privacy Policy explains the processing and retention considerations for searches and technical request data.
Known limitations
- Only the four named datasets are included; this is not a global collection of every sanctions, PEP or adverse-media source.
- The OFAC integration covers the SDN List and not every OFAC non-SDN list.
- Source data may be incomplete, approximate, delayed, corrected later or formatted unexpectedly.
- Name similarity produces false positives and can miss relevant records when names or transliterations differ substantially.
- Ownership and control, sectoral measures and unlisted affected parties require analysis outside a name-list search.
- The official authority and governing legal text always take precedence over the app.
Corrections and reproducibility
If a result appears inaccurate or outdated, record the source, official reference, searched phrase and the discrepancy, then email [email protected]. Do not send unnecessary identity documents or other sensitive information. Verify urgent or consequential cases directly with the responsible authority.
The appropriate use of Sanctions Sleuth is candidate discovery: find records worth checking, then verify identity, legal scope and current status through official material.