> ## Documentation Index
> Fetch the complete documentation index at: https://docs.alphaengine.trade/llms.txt
> Use this file to discover all available pages before exploring further.

# Privacy Model

> What AlphaEngine hides, what becomes public, what can leak, and what trust remains.

AlphaEngine privacy is strongest during submission and evaluation.

It should be described as private decisioning and private evaluation for public
markets, not as hidden public market execution.

## Required statement

<Callout>
  AlphaEngine keeps strategy inputs confidential during submission and evaluation,
  while scores, evidence artifacts, eligibility status, and final effects can
  become public.
</Callout>

## Privacy map

<AccordionGroup>
  <Accordion title="What is hidden">
    Private strategy parameters, protected component choices, pre-selection
    strategy logic, and private allocation or policy inputs when supported by
    the confidential path.
  </Accordion>

  <Accordion title="Hidden from whom">
    Public observers, competing builders, unauthorized operators, and anyone
    outside the configured evaluation boundary.
  </Accordion>

  <Accordion title="Protected phases">
    Submission, decoding and validation, simulation, scoring, and selection or
    eligibility review.
  </Accordion>

  <Accordion title="What becomes public">
    Scores, benchmark diagnostics, eligibility status, report hashes, artifact
    references, final execution effects, and market-level behavior after
    deployment.
  </Accordion>

  <Accordion title="What can still leak">
    Timing, metadata, supported market family, wallet or submitter information
    if revealed, final settlement effects, and behavior inferred from repeated
    public execution.
  </Accordion>

  <Accordion title="Remaining trust">
    Attested compute implementation, operator configuration, permission and key
    management, side-channel assumptions, artifact publication policy, and
    governance over upgrades and supported families.
  </Accordion>
</AccordionGroup>

## Related pages

<CardGroup cols={3}>
  <Card title="Confidential compute" icon="lock" href="/architecture/confidential-compute">
    Architecture-level compute boundary.
  </Card>

  <Card title="Trust and confidential compute" icon="shield" href="/trust/confidential-compute">
    Operational assumptions and limits.
  </Card>

  <Card title="Public settlement" icon="file-check" href="/trust/public-settlement">
    What becomes observable when effects settle.
  </Card>
</CardGroup>