Trust posture

Security for a private assistant workspace.

Neuron is built for assistant work that touches memory, files, integrations, schedules, approvals, and usage records. The controls below translate that security posture into plain language before customers connect data.

Memory visibility

Every assistant memory stays inspectable.

Plain English: Neuron should not quietly build a hidden profile that users cannot correct. Assistant memory is intended to stay visible, editable, and removable so people can see which context may influence future runs.

  • See what it remembers.Review the facts, preferences, files, and outcomes attached to an assistant.
  • Fix the record.Correct stale assumptions before they compound into future work.
  • Remove what no longer belongs.Delete memory that should not keep influencing the workspace.
Trust boundaries

Controls that keep assistant autonomy understandable.

Neuron can be powerful because it connects memory, tools, schedules, and outputs. These boundaries make that power legible instead of mysterious.

Integration boundaries

Connected services use explicit grants. Customers can disconnect integrations, and sensitive external writes are designed to stop for review before anything leaves the workspace.

Usage and billing audit

Spark usage, provider usage facts, checkout sessions, and billing events are reconciled so account balances can be explained without hand-waving.

Account controls

The product is planned around export, deletion readiness, credential-safe cleanup, rate limits, and abuse monitoring.

Sensitive work

Neuron slows down where a private assistant should.

The product is designed to translate advanced infrastructure into simple product behavior: ask before risky actions, show budgets before expensive work, and review larger capabilities before rollout.

Approval boundary

Sensitive external actions pause for review while lower-risk assistant work can continue inside account limits.

Provider routing

AI requests may use model and infrastructure providers for reliability, latency, cost, or capability.

Support review

Larger limits, custom tools, and team workflows can require manual review before they become available.

Trust boundaries

Clear controls for sensitive assistant work.

Payment details are handled by payment provider checkout and are not stored by Neuron. Assistant work stays behind account controls, integration grants, approval boundaries, and usage budgets.

Customers should not connect highly regulated data unless Feline Labs has approved that use case in writing. Written approval keeps data scope, provider routing, retention, and support expectations explicit.

Privacy details

The full Privacy Policy explains collected data, AI provider processing, retention, export, deletion, cookies, and customer choices.

Read privacy policy