• Secure computation

• Data in Use Protection

• Runtime Engine

Run the private part without exposing user details

Run the private part without exposing user details

StoffelVM runs `.stflb` bytecode across clear values, shares, client-provided values, and explicit openings
# vm_run.stfl
def main() -> int64:
  var private_value = ClientStore.take_share(0, 0)
  var adjusted = private_value.add_scalar(5)
  return adjusted.open()

What StoffelVM does

A VM for programs that mix ordinary values with secret shares

Keep private details as shares

Split sensitive user context into shares. Compute without pooling private details.

Approved results only

Open allowed output.

Built for real runs

Register-based architecture with deterministic execution, p50/p95/p99 observability, and protocol-aware scheduling. Designed for systems where performance characteristics need to be auditable, not estimated

Built for builder workflows

Separate clear and secret state

Keep public values clear and sensitive values represented as shares.

Deterministic execution

Same inputs. Same outputs. Easier debugging and review.

Deep observability

Trace the run, inspect latency, and debug the workflow.

How you run it

  1. Write in StoffelLang

Write the logic with public and secret types

  1. Compile to bytecode

Compile to `.stflb` with clear and secret values handled separately

  1. Execute with MPC

Compute over shares. Open only the result your application needs

© 2025 Stoffel Labs Inc. All rights reserved.

© 2025 Stoffel Labs Inc. All rights reserved.

© 2025 Stoffel Labs Inc. All rights reserved.