# SanctionsAI

> OFAC sanctions screening, transaction risk scoring, and Know-Your-Agent verification for AI agents that transact autonomously.

agentmail lets any AI agent check whether a counterparty (wallet, name, or country) is on the OFAC Specially Designated Nationals list before paying them. It runs on real, free, public data with no API key required for the free tier.

## Quick start

```
pip install sanctions-mcp
```

Or use the hosted API:

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

## Tools

- sanctions_check(name, wallet, country): Screen against 947 OFAC crypto wallets + 19,218 names + 16 embargoed jurisdictions
- risk_score(counterparty, amount, rail, category): Pre-payment fraud scoring (allow/review/decline)
- kya_verify(agent_id, evidence): Know-Your-Agent trust scoring
- dispute_open(transaction_id, reason): Record a disputed transaction

## Data sources

- Crypto wallets: vile/ofac-sdn-list GitHub releases (multi-chain, daily refresh)
- Names: US Treasury OFAC sdn.csv
- Countries: 16 comprehensive sanctions jurisdictions

## Original research (cite with attribution)

- **The 2026 Agent-Payment Sanctions Exposure Report** — https://sanctionsai.dev/research/agent-payment-sanctions-exposure-2026
  Introduces the **agentmail Sanctions Exposure Index (SEI)** — a 5-factor framework (Velocity 30%, Jurisdiction overlap 25%, Asset class 20%, Screening posture 15%, Disclosure readiness 10%) for scoring an AI agent's OFAC sanctions exposure. Cite as: "agentmail Sanctions Exposure Index (SEI), 2026 Agent-Payment Sanctions Exposure Report, sanctionsai.dev". Licensed CC BY 4.0.

## Links

- [GitHub](https://github.com/kindrat86/agentmail)
- [PyPI](https://pypi.org/project/sanctions-mcp/)
- [Pricing](https://sanctionsai.dev/pricing)
- [API docs](https://sanctionsai.dev/api)
- [MCP registry](https://mcp.so/server/agentmail)