
• MPC language
• Privacy by Design
Mark the sensitive part. Run it over shares. Open only what your product needs
What StoffelLang gives you
Python-style code for computations that mix public values with secret shares
Mark what must stay private
Use public and secret types. Make openings explicit. Catch the boundary before the run.
Keep the workflow familiar
Check, compile, and run locally before you wire the network.
Start from real patterns
Thresholds. Matching. Risk scoring. Federated aggregation. The private parts, written as code.
Why this is not another policy claim

Open only the result
Return the count, match, or score. Keep private details outside ordinary app storage.

Build with private primitives
Aggregates. Thresholds. Comparisons. Key operations. Write the useful part while marking the sensitive part.

Catch leaks before runtime
Make sensitive outputs explicit before they reach logs, dashboards, or incident reports
How you build with StoffelLang
Mark what's sensitive
`secret` for sensitive values. `public` for everything else.
Write your computation
Analytics, matching, risk scoring, and ML workflows — written like normal code.
Explicit reveals
Open the aggregate. Keep the records out of the result.
Your program opens the result while sensitive user context stays represented as shares.
