Synth Studio

An open-source platform that generates privacy-preserving synthetic data using machine learning. It solves a fundamental problem: organizations need to share, test with, and train models on sensitive data, but privacy regulations (GDPR, HIPAA) and internal policies block access.

THE PROBLEM: Data scientists can't access production datasets for ML training. Developers can't test with realistic data. Teams can't share data across departments or with contractors. The result: slower innovation, poor testing, and compliance bottlenecks.

THE SOLUTION: Synth Data Studio learns the statistical patterns in your data and generates new synthetic records that are mathematically proven to contain no real individuals. The synthetic data preserves correlations and distributions, so ML models trained on it perform comparably to those trained on real data.

TWO GENERATION MODES:

1. ML Training Mode: Upload a CSV, train a generative model (CTGAN, TVAE, or Gaussian Copula), and generate any number of synthetic rows. The platform learns your data's structure and creates statistically similar records.

2. Schema Mode: Define column types (name, email, SSN, credit card, etc.) and generate up to 1 million rows instantly. No training data required. Perfect for prototyping, demos, and cold-start scenarios.

PRIVACY & COMPLIANCE:

• Differential privacy with configurable epsilon/delta parameters
• One-click compliance reports (PDF) for HIPAA/GDPR audits
• Model cards showing exactly what the synthetic data contains
• Audit logs for enterprise governance

OPEN SOURCE & SELF-HOSTABLE: The entire platform is MIT licensed. Organizations can self-host on their own infrastructure for complete data sovereignty. No vendor lock-in, no data leaving your network.

TECH STACK: Python backend (FastAPI, SDV library), Next.js frontend, PostgreSQL database. Deployable via Docker Compose in minutes.

WHY NOW: AI regulations (EU AI Act, GDPR enforcement, HIPAA updates) are making synthetic data a compliance requirement, not just a nice-to-have. The market is shifting from "optional privacy tool" to "mandatory infrastructure."

Built by a student on a $0 budget using AWS Educate and GitHub Student Pack. Proving that enterprise-grade privacy tools don't require enterprise funding.