OpenSanctions is an international database of persons and companies of political, criminal, or economic interest.
External link for OpenSanctions
Berlin, DE
OpenSanctions reposted this
Vegas. ACAMS . #FollowTheMoney; literally. its also the name of the data model behind OpenSanctions, so the photo was non-negotiable. meeting partners and clients this week, and keen to meet new players too. dm me if you're here.
❗ ICYMI: We recently shipped a cool new toolkit for talking to our API — via your own code, the terminal, and MCP. Come and battle-test it with us👇
🔴 "Should compliance teams use LLMs in screening?” was a question we were asking just a few months ago. That future arrived much faster than we anticipated. LLM agents are no longer just reading lists, but running workflows: pulling customer records, running screening checks, flagging anomalies, and handing analysts a clean case file to sign off on. So now we’re asking a slightly different question: what exactly is your agent calling, and can it show its workings? yente-client is our Python-side tooling layer for talking to yente, our open-source screening API. It collapses several jobs that would ordinarily be handled separately into a single package: 🔧 SDK — for developers wiring screening into their own systems ⌨️ CLI — for analysts running ad-hoc checks or nightly re-screens from the terminal 🤖 MCP server — for LLM agents to handle screening as a step in a larger workflow Each layer offers a different way to call the API, but all three are powered by the same open matching engine. Whether a response arrives as CLI text, an SDK object, or an agent's tool result, it's a #FollowTheMoney entity underneath — a little slice of a bigger, connected graph, with a clear data trail. As compliance workflows keep shifting toward agentic tooling, we think one distinction will become increasingly important: not just that a match was flagged, but why. Come and battle-test yente-client for us! Links below 👇 #KYC #AML #sanctionsdata
🔴 "Should compliance teams use LLMs in screening?” was a question we were asking just a few months ago. That future arrived much faster than we anticipated. LLM agents are no longer just reading lists, but running workflows: pulling customer records, running screening checks, flagging anomalies, and handing analysts a clean case file to sign off on. So now we’re asking a slightly different question: what exactly is your agent calling, and can it show its workings? yente-client is our Python-side tooling layer for talking to yente, our open-source screening API. It collapses several jobs that would ordinarily be handled separately into a single package: 🔧 SDK — for developers wiring screening into their own systems ⌨️ CLI — for analysts running ad-hoc checks or nightly re-screens from the terminal 🤖 MCP server — for LLM agents to handle screening as a step in a larger workflow Each layer offers a different way to call the API, but all three are powered by the same open matching engine. Whether a response arrives as CLI text, an SDK object, or an agent's tool result, it's a #FollowTheMoney entity underneath — a little slice of a bigger, connected graph, with a clear data trail. As compliance workflows keep shifting toward agentic tooling, we think one distinction will become increasingly important: not just that a match was flagged, but why. Come and battle-test yente-client for us! Links below 👇 #KYC #AML #sanctionsdata
OpenSanctions reposted this
🪪 We collect, clean, and aggregate data from hundreds of sanctions lists, watchlists, and PEP registries worldwide. It can get pretty messy: each jurisdiction has its own approach to cramming nicknames, abbreviations, or transliterations into a single “Name” field in a table. Name-cleaning involves removing irrelevant text and punctuation, and correctly categorizing nicknames, former names, and transliterations. The task is embedded with cultural nuances as well as technical ones: naming conventions differ significantly across writing systems, languages, and regions. Our name-cleaning framework brings order to the chaos by providing a culturally aware approach to cleaning, while ensuring all entities in OpenSanctions speak a single #FollowTheMoney data language. Swipe for the highlights, or see our latest blog post below if you’re looking to dive deeper into the technicalities, including how we’re using LLMs to automate some of the cleaning pipeline 👇 #AML #KYC #sanctionsdata
✨ We’re hiring a part-time HR coordinator! We’re looking for a clear, warm communicator with hands-on experience owning recruitment and HR processes in similarly sized companies, supporting employees and freelancers across multiple countries. You’ll manage onboarding and offboarding, keep our HR processes organized, and develop new policies and practices. You’ll be savvy with contracts, records, and documentation, and able to work with a high degree of accuracy and confidentiality. ✨ Sound like you? Apply via the form on our website (linked below), where you can also find the full job description and requirements. We’ll ask for your CV, LinkedIn profile, and a few short answers in place of a cover letter 👇
🔎 “Has the data changed?” It's a question investigators, compliance teams, and journalists often need to answer. Our friend Stephen Abbott Pugh just shipped a cool new feature at OpenCheck called Watchlist, which lets you track a company and get notified when its record changes. It works by watching the delta files — including the one OpenSanctions publishes several times a day — that list every entity added, modified, or removed. When a watched company shows up in that delta, OpenCheck triggers a fresh check. More in comments 👇 #AML #KYC #LEI #opendata
OpenSanctions reposted this
‼️ ICYMI: we're starting a series to showcase what OpenAleph can do and why it's an absolute must for any investigator working with lots of data. In the first post, we follow the trail of a sanctioned vessel, the Ji Song 8, using records from several datasets provided by OpenSanctions. One IMO number leads to a company. A company leads to more vessels. A shared address leads to more companies. The fun goes on and on. It’s a good example of one of the perks of using OpenAleph: when you combine a number of different data sources, you don't always know where they'll take you. You can follow the identifiers and relationships you discover along the way, all while preserving your research trail and the underlying sources that got you there.
🚢 Who owns the Ji Song 8, and what else do they own? Our friends over at the Data and Research Center have launched a new series that walks you through how to use #OpenAleph in investigations. Their first mini investigation uses several OpenSanctions data sources — you’ll learn how to use OpenAleph to find connections, inspect the underlying data, and decide where to go next. Link in comments! 👇
Summer’s essentially over (sorry!). Time to get back to the good stuff. ✨ We’re launching a new series to show what OpenAleph can do and why it belongs in every investigator’s toolkit. First up: a mini investigation into a sanctioned vessel, a few different datasets courtesy of OpenSanctions, and a long trail of connections. We start with a simple question: Who owns the Ji Song 8? From there, we follow IMO numbers, company names and shared addresses to uncover a wider network. The goal is to show what OpenAleph does best: helping you follow the trail across datasets, keeping the underlying evidence in view, and giving you the freedom to decide what’s worth checking out next. 🔎 Read the first post in the comments below. 🔎 Want to try OpenAleph for yourself? Get started on our public instance: https://lnkd.in/gPECKA7s 🔎 Or, if you’re interested in having OpenAleph hosted and managed for your organization, learn more here: https://lnkd.in/ginssGJp
OpenSanctions reposted this
��� In November 2023, a company was registered in Kursk, Russia. It doesn’t appear on any sanctions lists, and yet it’s linked to a sanctioned individual. The Russian company register indicates that one of its major shareholders is a hotel management company, half of which is owned by someone Ukraine designated thirteen months earlier. Our data has always carried that chain, but our new risk tag, sanction.control, now separates ownership from adjacency: sanction.linked ➡️ Direct relationships to a sanction or a sanction.control entity sanction.control ➡️ Companies and assets that are direct or indirect subsidiaries of sanctioned entities. Any stake, any number of layers down. Our latest blog post dives into the mechanics — link in comments 👇 📷 Timon Studler via Unsplash
🪪 We collect, clean, and aggregate data from hundreds of sanctions lists, watchlists, and PEP registries worldwide. It can get pretty messy: each jurisdiction has its own approach to cramming nicknames, abbreviations, or transliterations into a single “Name” field in a table. Name-cleaning involves removing irrelevant text and punctuation, and correctly categorizing nicknames, former names, and transliterations. The task is embedded with cultural nuances as well as technical ones: naming conventions differ significantly across writing systems, languages, and regions. Our name-cleaning framework brings order to the chaos by providing a culturally aware approach to cleaning, while ensuring all entities in OpenSanctions speak a single #FollowTheMoney data language. Swipe for the highlights, or see our latest blog post below if you’re looking to dive deeper into the technicalities, including how we’re using LLMs to automate some of the cleaning pipeline 👇 #AML #KYC #sanctionsdata