Synthetic data · DataCebo, Inc. (open-source project, originated at MIT)

Synthetic Data Vault

Open-source Python library for generating synthetic tabular, relational and time-series data from real datasets you run yourself.

Visit Synthetic Data Vault Alternatives

The Synthetic Data Vault (SDV) is a Python library, not a hosted product: it provides a set of generative models (Gaussian copulas, CTGAN and other deep-learning synthesizers) that learn a real dataset's structure and sample new synthetic rows from it, supporting single tables, multi-table relational schemas and time-series data. Because it runs entirely inside a team's own environment, production data never has to be sent to a third party — the privacy guarantee comes from local execution plus a companion SDMetrics library that scores how well synthetic data preserves statistical properties without directly matching real records. It originated as an MIT research project and is maintained by DataCebo, which also sells a paid, hosted enterprise version with additional privacy and compliance features on top of the open-source core.

At a glance

Vendor DataCebo, Inc. (open-source project, originated at MIT)
Pricing model Open source + paid options
Free tier Yes
Deployment Self-hosted
Open source Yes (BSL-1.1)
Best for Data scientists who want a self-hosted, code-first synthetic data library rather than a managed platform.

Pricing

The core library is free and open source under a source-available license; DataCebo sells a separate paid enterprise product.

Pricing has not been verified yet — see the vendor's site.

Features

  • Single-table, multi-table relational and time-series synthesizers
  • Gaussian copula and deep-learning (CTGAN) generation models
  • SDMetrics companion library for fidelity and privacy scoring
  • Constraint definitions to preserve business rules in generated data
  • Runs entirely in the user's own environment
  • Python API for pipeline integration
  • Active open-source community and research lineage from MIT

Integrations

Profile last reviewed September 21, 2026

Head to head

Alternatives

MOSTLY AI Synthetic data platform that generates statistically representative tabular/relational data with built-in differential privacy. Tonic.ai Suite of synthetic/de-identified test-data tools (Structural, Textual, Fabricate) for databases, unstructured text and mock data. Synthesized Synthetic and test-data platform for generating privacy-preserving, statistically representative data to develop and validate ML models.

Synthetic Data Vault in the index now

Terms to know

Machine learning Building models that learn patterns from data to make predictions or decisions, rather than following hand-coded rules. Data vault A modeling methodology that splits data into hubs, links, and satellites to make a warehouse auditable and adaptable to change.

Related guides

How to choose a synthetic data tool Decide what kind of data you need synthesized, where generation must happen, and whether an open-source library or a vendor platform fits your compliance bar.