OpenSanctions

OpenSanctions consolidates the world's sanctions lists, government watchlists, and registers of politically exposed persons, meaning people in prominent public roles who carry a higher corruption risk, into a single clean, standardised dataset. It draws from hundreds of official sources, among them the US, EU, UK, UN, and many national agencies, and models every entity with one shared schema called FollowTheMoney, so a person, company, or vessel looks the same regardless of which list it came from.

Everything on the free tier is available for bulk download: entity data as JSON, plus CSV and other targeted exports, refreshed daily so it stays current. There is also a matching and reconciliation API, hosted or self-hosted, for screening a given name against the data. For RAG the bulk JSON is usually the entry point: store each entity with its properties and source references, and generate a short natural-language summary per entity so your retriever has readable text to match against.

It is built for compliance, anti-money-laundering, and due-diligence work: a screening assistant, a risk-research tool, or a knowledge graph that joins who is sanctioned to who controls what. Because it links entities to OpenCorporates wherever they overlap, you can combine sanctions data with corporate ownership in a single graph.

The free tier is CC BY-NC 4.0, so the NonCommercial term applies: building it into a commercial product needs the separate commercial licence, and attribution is required either way. Treat the data as a lead rather than a verdict. It aggregates source lists of varying quality, name matching produces false positives, and none of it is legal advice, so always trace a hit back to its original source before acting on it.

It complements OpenCorporates directly and overlaps in places with Wikidata's PEP coverage and GLEIF identifiers, but it is the most focused, compliance-ready option of that group.