Getting permission to use health data is slow. Subsalt makes it faster.
Create a de-identified digital twin of your data and run analyses securely against original records without ever exposing PHI.
Coming to a Breaking Point
Data access review can't scale.
The governance process was built for dozens of data access requests a year — reviewed one by one in committee. Your teams now submit hundreds: researchers wait 6 to 18 months for data you already own. AI agents will submit thousands, at machine speed. No committee process can scale to that.
How It Works
Synthetic data for exploration. A secure runtime for production.
Start with synthetic data — no IRB, no BAA, no wait. Move to the runtime when you need real-data results. Every output is audited before release.
Synthetic Data
Statistically faithful synthetic datasets, de-identified and attested. Available in minutes.
What disappears
- IRB review
- BAA negotiation
- Restricted data use agreements
- Governance review for sharing
Secure Runtime
Run code against real patient data. Every output verified de-identified before release.
Request early accessWhat you get
- Publication-ready results
- Auditable data lineage
- Validates synthetic-trained models
- Compliance artifact on every result
Proven in Production
Trusted where the stakes are highest.
See what your data can do.
We benchmark synthetic data against your real data, on your use cases, in your analytical workflows.
Schedule a Proof of ValueBuilding for AI? The compliance infrastructure for research is the same infrastructure AI needs at scale.