Case Study
LimitGuard
Compliance scoring API dual-exposed as an MCP server — auditable jurisdictional risk data that AI agents can call directly.
Metrics
- 185
- jurisdictions scored
- 0–20
- risk score range
- Dual
- MCP + REST exposure
- mTLS
- transport security
Problem
Give AI agents a callable, auditable risk-scoring tool they can trust.
Stack
- FastAPI scoring 185 jurisdictions (CPI index + FATF grey/black lists) into a 0-20 risk score
- mTLS
- Tenant isolation
- Dual-exposed as an MCP server (AI agents can call it directly)
- Semgrep + OWASP ZAP in CI
Constraints
Compliance primitive used by downstream AI agents; must be independently auditable.
Outcome
A compliance primitive that AI agents can call directly, with jurisdictional risk data as a first-class API.
Want something like this?
Start a project