Resources
Sanctions screening resources: tools and research
Find a sanctions screening quickstart, tools for coding agents, and original research. Choose a resource for the integration or review task in front of you.
Sanctions screening API quickstart: your first request
Create a live API key, choose an available source, and run your first check with your free lifetime allowance. Sandbox testing is optional.
Explore the resourceMCP server: overview and connection options
See which coding agents can connect, what the tools do, and how sign-in works.
Explore MCP connection optionsMCP integration: step-by-step walkthrough
Add the server, authenticate, run a synthetic request, and troubleshoot the integration.
Follow the MCP walkthroughSanctions-data research: dates, names and identifiers
Examine two official snapshots, field definitions, and limitations. Includes aggregate data and reproduction files.
Read the research and methodologyThe OFAC List Said Yes. Google AI Said No.
A 96-company sanctions screening study found 17 incorrect Google AI answers about OFAC SDN List membership, including three listed companies described as absent.
Read the AI study and evidenceBuild with the tools you already use
Start with the task you want to finish. To make your first request, follow the API quickstart. To work with a coding agent, connect the SanctionsKit MCP server. The agent can read the docs, check a request, and test the integration with invented sandbox data.
The MCP integration guide covers setup and a starting prompt. For exact request and response fields, use the API reference or download the OpenAPI specification. Your deployed application uses the REST API whether you write the integration yourself or use an agent to help.
Keep the result connected to its scope
A useful integration preserves the selected sources, screening status, and evidence references. Read the practical guides to understand name matching, identity review, and the limits of a list check. Use the integration checklist to cover retries, unavailable sources, retention, and the path from a potential match to a human review.
Sandbox examples use invented subjects and records. Check current coverage separately when planning production source selection. A successful example shows how a request works; it does not establish that a real subject has no applicable restrictions.
Common questions
Where should I start if I want to add sanctions screening to an application?
Use the API quickstart to create a live key, choose an available source and make one request. Use the optional sandbox for made-up test records. Then choose a language guide or connect an AI coding agent through MCP. The API reference and OpenAPI document show the current fields and actions.
Which resources help a reviewer understand screening evidence?
The practical guides explain name matching, identity review, and source limits. Original research examines specific source fields or screening questions and provides its method and evidence. Use those findings within their stated scope rather than treating them as a result for your own customer.
Are the public demo and sandbox examples real sanctions checks?
No. They use invented subjects and records to show the workflow. A successful example demonstrates request and review behavior. Check production coverage separately before screening real information under your organization’s approved process.
Do I need an AI coding agent to use the SanctionsKit API?
No. You can use an SDK, follow a complete HTTP example, or build a client from the OpenAPI contract. MCP is another way to help write and test your integration. The deployed application still calls the same REST API from its backend.