Skip to main content

Introduction

See it live

Agent Disaster Lab → — watch AI agents cause real regulatory disasters across 9 scenarios (fintech, healthcare, insurance, data), then watch comply54 block every one. Demo mode requires no API key.

What is Comply54?​

Comply54 is the African AI compliance infrastructure layer. It enforces African regulatory requirements — NDPA 2023, CBN transaction controls, KDPA 2019, POPIA, and 13 other African data protection laws — directly inside your AI agent stack.

One import. Zero Rego knowledge required.

from comply54.sectors import NigeriaFintechCompliance

compliance = NigeriaFintechCompliance()
result = compliance.check(
    action="transfer_funds",
    params={"amount": 15_000_000, "currency": "NGN"},
    context={"kyc_tier": 3},
)

if result.blocked:
    raise ValueError(result.primary_violation.messages[0])
# → "CBN NIP Framework: Transaction of ₦15,000,000 exceeds
#    ₦10,000,000 single-transaction cap — blocked"

Why Comply54?​

African AI companies operating in regulated sectors face a compliance problem that generic tools don't solve:

  • Presidio and spaCy have no awareness of BVN, NIN, Ghana Card, KRA PIN, or any African identity format
  • GDPR/CCPA tools don't cover NDPA 2023, CBN AML/CFT regulations, or KDPA cross-border transfer restrictions
  • OPA is powerful but requires every team to author and maintain their own Rego policies
  • No existing tool enforces CBN transaction limits (₦10M NIP cap), NFIU CTR thresholds, or NDPA §25 cross-border rules at agent runtime

Comply54 solves this at the enforcement layer — the point where an AI agent is about to take an action. It evaluates the action against the applicable regulations and returns a structured decision before anything is executed.

How it works​

Agent decides to take an action
           │
           ▼
  comply54.check(action, params)
           │
           ├─ NDPA 2023:       cross-border? data type? consent?
           ├─ CBN Controls:    amount within limits? KYC tier?
           ├─ BVN/NIN:         biometric data in output?
           ├─ NFIU AML:        CTR threshold? suspicious pattern?
           └─ OWASP Safety:    PII leak? prompt injection?
           │
           ▼
    PolicyDecision
    ┌─────────────────────────────────────────────────────────┐
    │ overall:        "deny"                                  │
    │ regulation:     "CBN NIP Framework"                     │
    │ message:        "₦15M exceeds cap"                      │
    │ rule_triggered: "nip_cap"                               │
    │ citations:      [CBN NIP Framework §4.2 — Tx Cap]       │
    │ audit_id:       "aud_abc123"                            │
    └─────────────────────────────────────────────────────────┘

Key features​

FeatureDetails
No OPA requiredBuilt on regopy — evaluates Rego policies in-process. No binary to install, works in Lambda and Edge functions.
21 policy packsNDPA, CBN, NHA 2014, NAICOM, KDPA, POPIA, Ghana DPA, Rwanda DPA, Egypt PDPL and 12 more — with per-rule regulatory citations on every decision.
Sector packsNigeriaFintechCompliance, NigeriaHealthcareCompliance, NigeriaInsuranceCompliance, KenyaFintechCompliance, PanAfricanFintechCompliance — pre-composed for your vertical.
Framework adaptersNative integration with LangChain, LangGraph, CrewAI, AutoGen, and Vercel Eve via comply54-adapter-eve.

Regulatory coverage​

RegionFrameworks
NigeriaNDPA 2023, CBN Transaction Controls, BVN/NIN Framework (NIMC Act 2026), NFIU/MLPPA 2022, National Health Act 2014, Insurance Act 2003 / NAICOM
KenyaKenya Data Protection Act 2019
South AfricaPOPIA (Protection of Personal Information Act)
GhanaGhana Data Protection Act 2012
RwandaLaw No. 058/2021 on Personal Data Protection
EgyptPersonal Data Protection Law No. 151/2020
EthiopiaPersonal Data Protection Proclamation 1321/2024
MauritiusData Protection Act 2017
TanzaniaPersonal Data Protection Act 2022
UgandaData Protection and Privacy Act 2019
UniversalOWASP Agentic AI LLM01, LLM06, LLM08, LLM09

Next steps​