Could you direct me to the API documentation for getting the data of Sanctions Programs, Associated Entitles, Datasets? Also may I know if there a formula or logic to calculate the confidence%? Thanks in advance!
1 Like
Hey there!
Thanks for the thoughtful question. Happy to get you up and running with some docs and info! 
A great place to start here would be with our API overview. For the API documentation, you can check out https://api.opensanctions.org and https://api.opensanctions.org/docs.
For sanctions programs, check out our docs programs and programIds: https://www.opensanctions.org/faq/168/programs/. A good intro to program-mapping is here in this article we recently published as well: https://www.opensanctions.org/articles/2025-07-09-program-mapping/.
For entities and associated entities, the Entities guide explains schemas and properties. The Statements guide shows the underlying model and how relationships link people, companies, vessels, and more: https://www.opensanctions.org/docs/statements/. Our Data dictionary is here as well: Data dictionary - OpenSanctions
On scoring, the best starting points are:
• Configuring the Scoring System
• Fine-Tuning the Scoring
The scores you see from the /match API are designed to order candidates by how well they fit the input, not to give an absolute truth. A score of 0.9 doesn’t mean there is a 90% chance of a match. It only means the system thinks that candidate is stronger than one with 0.7.
The most reliable way to use scores is to calibrate thresholds against your own data. Take a sample of your portfolio, run it through the matcher, and see where the true positives and false positives fall. Then pick a cutoff for “matches” that balances the precision and recall you need. For some use cases that might be 0.8, for others you may want to lower or raise it.
Let me know if you have any specific questions I can help you out with, and I’ll be happy to explain anything you need clarification on
2 Likes