Synthetic clinical data. Evidence you can review.
Turn a clinical CSV into synthetic data for prototypes, analysis workflows and research evaluation. Review measured privacy and utility checks before deciding how to use the result.
Synthetic CSV + downloadable evaluation PDF. Your table is generated in the EU (Frankfurt).
From a de-identified table to evidence you can inspect
Upload
A de-identified clinical CSV. We start from the table you bring.
Intelligence OMOP
Intelligence maps each field to OMOP when the dictionary agrees — labs stay labs, flags stay flags.
Intelligence Generate
Our 5-level intelligence system generates, scores, and retunes against privacy and utility.
Download the evidence
A synthetic CSV and an evaluation PDF.
Choose a plan for your current workload
Paid plans renew automatically each month. A run buys generation and evaluation, not a favorable result.
Explore
One evaluation to try generation and the report
- 1 evaluation total
- 1 active project · 200–1,000 rows
- Gaussian copula
- CSV and evaluation report
Starter
For one project that you evaluate repeatedly
- 5 evaluations per month
- 1 active project · 200–5,000 rows
- Gaussian copula
- CSV and evaluation report
Professional
For teams running several dataset evaluations
- 20 evaluations per month
- 5 active projects · 200–5,000 rows
- Gaussian copula
- CSV and evaluation report
Prices exclude applicable taxes. Stripe Checkout shows the payable total before you confirm.
No trial, rollover or automatic overage. Custom work is available via contact, not a fourth plan.
Discuss custom workBefore you start
Does synthetic mean anonymous?
Not automatically. Generation can reduce reliance on original records, but residual disclosure risks and usefulness must be assessed for each dataset and purpose.
Does the report certify compliance?
No. The evaluation report records measured checks and limitations. It is not GDPR or HIPAA certification, legal approval or clinical validation.
Can I deploy Gesalp on-premise?
The current offer is a hosted web application. On-premise deployment is planned and is not included in the self-service plans.
What can I upload?
Use one flat CSV with consistent column names and values. Real de-identified records are allowed. Check your plan’s row limits and the 10 MB file size. There is no column cap on current plans.
Where is my table generated?
Your table is generated in the EU (AWS Frankfurt).
What if a check fails or cannot run?
Review warnings and failures before using or sharing the output. Unavailable checks are marked N/A with a reason; missing evidence is not treated as a pass.
Roadmap
What we’re working toward
How it works is live. These cards are next — not included in today’s plans.
Sharper OMOP maps
Every upload already maps columns to OMOP when the dictionary agrees. Next: fewer REVIEW fields, with human review still in the loop. Not a complete OMOP data-model export.
Tighter privacy and utility
The 5-level system already generates, scores, and retunes. Next: stronger trade-offs across more tables. Passing every check is not guaranteed.
On-premise deployment
Live synthesis runs in Frankfurt. Next we are exploring customer-managed infrastructure. Availability and scope are not committed.