How to Set Up OFAC Sanctions Screening for Your AI Agent

A step-by-step guide to adding OFAC sanctions screening to any AI agent using HTTP, MCP, or CLI. Zero configuration required for the free tier.

Adding sanctions screening to your AI agent takes one API call and under five minutes. This guide covers all three integration methods: HTTP API, MCP tool, and CLI.

Method 1: HTTP API (simplest, works with any agent framework)

Call our HTTP endpoint with the wallet address, name, or country you want to check. No API key needed for the free tier (5 checks/day per IP).

curl "https://sanctionsai.dev/sanctions?wallet=0x098B716B8Aaf21512996dC57EB0615e2383E2f96"

Response includes clean (boolean), matches (array of matched SDN entries), and quota (remaining free checks). Wire this before your agent’s payment function.

Method 2: MCP tool (works with Claude Code, Cursor, Windsurf)

Install pip install sanctions-mcp, configure the MCP server, and your agent can call sanctions_check as a native tool.

pip install sanctions-mcp
# Add to your MCP config:
# mcpServers.sanctions.command = "python -m sanctions_mcp"

Method 3: CLI (for scripts and cron jobs)

sanctions check --wallet 0x098B716B8Aaf21512996dC57EB0615e2383E2f96

What to do with the result

If clean=True, the transaction is safe to proceed. If clean=False, halt the transaction, log the match, and alert a human. Every check is timestamped for your compliance audit trail.