DataCebo SDV Enterprise

DataCebo, Inc. (Boston, MA, per its license agreement) builds the Synthetic Data Vault (SDV), a source-available Python library, and the commercial SDV Enterprise for synthetic tabular and relational data. The company's announcements describe a 2023-12 launch with 8.5 million USD in seed funding. The founding year is not stated on the pages reviewed.

SDV Enterprise runs inside the customer's environment and models relationships, business rules and statistical patterns across multi-table databases. Companion tools include SDMetrics for evaluation and SDGym for benchmarking, plus bundles for constraint-augmented generation and differential privacy. SDV 2.0, announced 2026-09, introduces generative relational models. Named users include ING Belgium, MAPFRE and Spar Nord.

• Generative relational models: Model an entire multi-table database in one model.
• SDV Enterprise: Commercial version deployed in the customer's environment.
• Constraint-augmented generation: Enforce business rules in synthetic output.
• Differential privacy: Train synthesizers with privacy guarantees.
• SDMetrics and SDGym: Evaluate and benchmark synthetic data quality.
• AI Connectors: Infer metadata and exchange data with databases.

A bank can generate synthetic transaction databases to train fraud models. An insurer can build realistic claims data for analytics vendors. An IT team can create multi-table test data for ERP upgrades.

Keeps data and models inside the customer's security boundary. Models multi-table relationships and business rules, not just single tables. Case studies report 31 percent more fraud cases caught at MAPFRE and 100x test coverage at ING Belgium (company-reported). A free community edition lets teams evaluate first.

Enterprises need synthetic data that preserves relational structure for software testing, fraud detection and AI evaluation. Vietnamese banks and insurers with large relational core systems can apply it to testing and analytics under data protection limits.

DataCebo, Inc. Launched 2023-12 with 8.5 million USD seed funding (company announcement). 2025-06: 10 million SDV downloads and differentially private synthesizers. SDV Community uses a business source license.