USE CASES

Synthetic financial data that behaves like the real thing

Balanced ledgers, sub-ledgers, master data, document flows, and ERP export formats — with ground-truth fraud labels. A naive generator fails the first sanity check; VynFi's data holds up to an auditor's eye and a fraud model's training run.

Why the structure matters

Most "synthetic data" is random rows that look right in a spreadsheet and fall apart the moment you check them: entries that don't balance, amounts that violate Benford's law, no document flow linking an invoice to its payment to its ledger posting. Real audit tooling, ERP pipelines, and fraud-detection models see straight through that.

VynFi generates behaviorally-faithful data instead — double-entry that balances, document-flow graphs, Benford-conforming amounts, and ground-truth fraud labels — built on the open-source DataSynth engine (Apache 2.0, 100k+ rows/sec).

Frequently asked questions

What makes VynFi different from a general synthetic-data tool?

General tools give column-level plausibility. VynFi generates structurally-valid financial data — double-entry-balanced ledgers, sub-ledgers that reconcile to the GL, document-flow graphs linking invoices to payments to postings, Benford-conforming amounts, and ERP export formats (SAP, SAF-T) — plus ground-truth fraud labels. That structure is what makes the data usable for audit analytics, ERP testing, and fraud ML.

Is the engine open source?

Yes. The DataSynth generation engine is open source (Apache 2.0) and runs at 100k+ rows/sec. The commercial platform adds the hosted API, scale, and the ERP/audit output formats.

How does pricing work?

Pure prepaid. 5,000 free non-expiring credits on signup (no card), then one-time credit packs from $19. One credit generates one row. No subscription, and every feature is open on every account.

What output formats are supported?

CSV, JSON, and Parquet for the core data, plus an SAP integration pack (BKPF/BSEG/ACDOCA) and SAF-T XML for five OECD jurisdictions, and the document-flow graph linking entries.