Skip to documentation

Data and security

Local-first is an ownership boundary, not a marketing adjective.

The Mac is intended to hold the canonical project. Local work should not require a cloud account, and future remote compute must expose data movement, region, cost, credentials, and retention before a job begins.

Status: design contract; product under active development

Reviewed:

How to read this page

This page defines the intended product boundary; it does not claim that a public application or cloud service is operating today. Implemented controls, tested release behavior, deployed-service controls, and independent assurance will remain separate evidence classes. Tensor Cortex does not currently claim SOC 2, ISO 27001, a completed penetration test, or third-party security certification.

The local project is authoritative

  • Creating and running a local project should not require a Tensor Cortex account.
  • Dataset sources, cleaned revisions, model files, checkpoints, eval results, and packages remain on the Mac unless the user explicitly exports or submits them.
  • Automatic fixes create new revisions rather than silently rewriting source data.
  • Project manifests should avoid embedding credentials and should keep machine-specific secrets outside portable artifacts.
  • Optional telemetry is planned as opt-in and must exclude datasets, prompts, outputs, model weights, and secrets.

Major data classes and intended defaults

Intended default handling for Tensor Cortex Studio data
DataDefault locationBoundary
Source and prepared datasetsLocal projectNo cloud transfer without an explicit job or export action.
Base models and adaptersLocal model storageLicense, source, revision, and checksum should travel with provenance.
Experiment configurationPortable manifestConfiguration is versioned; secrets and provider credentials stay outside the manifest.
Checkpoints and resultsLocal artifactsRetention and deletion remain under the project owner’s control.
Evaluation recordsLocal evidenceScores, cases, hardware, and known regressions are attached to the experiment.
Website contact emailEmail service providerHandled separately under the public Privacy Policy.

Future cloud compute must be explicit

When managed cloud jobs become available, the submission boundary is intended to show:

  • The provider or service class, region, hardware class, and data residency implications.
  • Which dataset revision, model revision, recipe, and artifacts will be transferred.
  • An estimated duration and price range, plus a hard maximum budget.
  • Checkpoint, retry, interruption, refund, retention, and deletion behavior.
  • Whether credentials belong to Tensor Cortex Cloud or a user-controlled BYOC account.
No silent fallback: local execution should never become a cloud job merely because the Mac is slow or lacks memory. The user must choose the remote boundary.

Models, datasets, and third parties

An open-model name does not settle license, retention, training-data, export-control, or commercial-use questions. Tensor Cortex Studio is intended to record source and license metadata, but users remain responsible for confirming that their dataset, base model, teacher model, and output use are permitted.

Report a security issue

Send the affected public surface, reproduction steps, and potential impact. Do not include live secrets, private datasets, or model credentials.

security@tensorcortex.com