Synth Studio is an open‑source, browser‑based tool that generates synthetic CSV or JSON datasets while preserving statistical properties and enforcing differential privacy. It automatically detects PII, offers model‑based or schema‑based generation, and provides quality metrics and compliance documentation for regulated data sharing.
Funding
Funding not disclosed
Founders
Product
Problem
Organizations that need to share data for analytics, testing, or collaboration often face legal and privacy constraints such as GDPR and HIPAA, which prohibit exposing real personal records. Traditional anonymization methods can either leave re‑identification risks or destroy the statistical utility of the data.
Solution
Synth Studio is an open‑source, browser‑based platform that creates synthetic datasets from CSV or JSON inputs while preserving the underlying statistical patterns. The tool automatically detects personally identifiable information and applies differential privacy with configurable budgets to ensure that generated records cannot be traced back to real individuals. Users can choose between training advanced generative models (CTGAN, TVAE) or using fast schema‑based generators, both supporting privacy‑preserving options. After generation, the platform provides quantitative quality metrics and audit trails, and exports compliance‑ready documentation for secure data sharing.
Target Audience
Primary users are data scientists, compliance officers, and developers in regulated industries such as healthcare, finance, and insurance who need to share realistic data without exposing real personal information.
Features
- Automatic schema detection and PII flagging for uploaded CSV/JSON files
- Differential privacy enforcement with adjustable privacy budgets
- Choice of model‑based generators (CTGAN, TVAE) or instant schema‑based generation
- Built‑in quality evaluation scores comparing synthetic data to the source
- Full audit trail and compliance PDF export for GDPR/HIPAA documentation
- Browser‑only operation with no data leaving the user’s environment; optional self‑hosting
- Open‑source MIT license with free usage and community contributions