Early accessSome features may be unavailable
AML Compliance

AML Test Data

Generate labeled transaction data with 14 money-laundering typologies, multi-party entity networks, and ground-truth labels for every record.

14 AML Typologies with Ground Truth

Each typology generates realistic transaction patterns calibrated against FinCEN enforcement actions

1Structuring / Smurfing
2Layering through shell companies
3Round-tripping
4Trade-based money laundering
5Funnel accounts
6Rapid movement (pass-through)
7Foreign exchange manipulation
8Nested correspondent banking
9Payable-through accounts
10Crypto on/off ramp
11Real estate integration
12Cash-intensive business mixing
13Loan-back schemes
14Insurance policy manipulation

Built for Compliance Teams

Multi-Party Networks

Each typology generates realistic entity networks: beneficial owners, shell companies, correspondent banks, and intermediaries with proper jurisdiction assignments.

Velocity Features

Transaction velocity, amount distribution shifts, counterparty diversity, and time-of-day patterns are calibrated against real FinCEN enforcement actions.

Sanctions Screening

Generate entities that match OFAC SDN, EU sanctions, and UN lists at configurable fuzzy-match thresholds for testing screening system recall.

VynFi vs Alternatives

FeatureManual CasesProd SamplingGeneric SynthVynFi
Ground-truth labels
Typology coverage
Network structure
Scale (millions of txns)
Privacy safe
Cost

Python SDK Example

import vynfi

client = vynfi.Client(api_key="vf_live_...")

job = client.generate(
    sector="banking",
    module="aml",
    typologies=["structuring", "layering", "round_tripping"],
    rows=500_000,
    suspicion_rate=0.05,     # 5% suspicious transactions
    network_depth=3,         # 3-hop entity networks
    output_format="parquet",
)

# Download with ground-truth labels
df = job.to_pandas()
print(df[df["is_suspicious"]].groupby("typology_id").size())

Stop waiting for production data samples

Generate labeled AML test data in minutes. 5,000 free credits to start.