Pricing
Start free. Pay as your pipelines grow. No surprise overages.
- CSV and Parquet export
- Up to 3 schemas
- Fidelity scoring on every job
- Community support
- Single user
- Everything in Sandbox
- API access + Python/R clients
- Up to 20 schemas
- Differential privacy controls (epsilon config)
- Priority email support
- 5 team seats
- Job history + audit log
- Everything in Growth
- Unlimited schemas and seats
- On-prem or VPC deployment option
- Custom SLA and uptime guarantee
- Dedicated onboarding
- Membership inference audit report
- Procurement and legal review support
Common questions
No. We fit a statistical model over the schema and any sample you provide. No raw rows are written to our systems. What we retain is a set of distribution parameters and correlation structure, not records.
One synthetic row in any output dataset. Schema connections and model-fitting jobs do not count against your monthly limit, only generated output rows.
Yes. Upgrades take effect immediately and you're billed pro-rated for the remainder of the current billing period. Downgrades take effect at the next billing cycle.
Three components: Jensen-Shannon divergence per column marginal (are individual column distributions preserved?), pairwise correlation matrix distance (are relationships between columns preserved?), and a train-on-synthetic test-on-real accuracy delta on a held-out classifier. All three ship with every generation job.
No. Synthpylon is purpose-built for tabular structured data: relational schemas, flat CSVs, columnar formats. We do not support free-text, images, or audio synthesis. Time-series with temporal dependencies is on our roadmap.
Not yet. Synthpylon is built with SOC 2 controls in mind and we plan to pursue certification as the company matures. If your procurement process requires a formal certification today, reach out and we can share our current security posture documentation.