See what Daymark can use, save, and keep out.

Do not enter or send confidential, client, employer, personal-sensitive, regulated, proprietary, repository, medical, legal, financial, student, or child-related material.

Do not enter credentials, secrets, tokens, cookies, claims, account IDs, session details, env values, DB URLs, Clerk keys, allowlists, or raw row contents.

Do not paste source documents, raw excerpts, contracts, production records, private examples, or account/session/auth material.

Use only high-level non-sensitive summaries and approved workspace categories until a separate sensitive-data/legal/privacy gate is accepted.

Browser-local personalization changes display preferences only; there is no hidden learning and no server-side personalization.

Owner memory inspection is owner-only and shows safe summaries, statuses, and categories rather than identities or row contents.

The Model-Ready Packet previews only; the server-side model scaffold is disabled, there is no live model call, and there is no send button.

Support, revocation, and incident handling are manual, not automated.

Access changes require human review outside this page; this route does not grant or revoke access.

Incidents, privacy concerns, and cohort revocation requests must be handled by the operator before any production-scope change.

This page is orientation copy only and does not create support workflow automation.

Public-safe means content may appear on public surfaces only when approved and non-sensitive.

Internal means content may support internal work but should not be published or exposed externally.

Restricted means content requires strict handling and should not be used in public, external, or broad AI contexts.

Review Required means privacy, sensitivity, ownership, and AI eligibility must be checked before use.

These labels are static orientation copy; they do not classify submitted material, ingest sensitive content, or approve privacy posture.

Review Required preserves no raw sensitive-data intake, no client workflow, no uploads, no integrations, no open access, and no access expansion.

AI eligibility remains unapproved until a separate human review checks source authority, ownership, sensitivity, and use boundary.

Human approval is required before privacy posture changes, AI eligibility approval, public claims, legal/privacy/commercial changes, sensitive-material intake, live data use, integrations, measurement, billing, file intake, backend behavior, or production-scope expansion.

Review means notes, sources, privacy and AI eligibility, and scope boundaries must be checked before use.

Hold means missing privacy classification, ownership, approval, source authority, or scope clarity blocks progress.

A separate implementation packet is required for changes outside copy/docs orientation; this page does not approve them.

Copy orientation may clarify privacy labels, AI eligibility meaning, route copy, and approved-material documentation without changing behavior.

Sensitive-data handling, privacy posture, AI eligibility approval, connected systems, file intake, measurement, billing, or live workflow require a separate packet.

Implementation packets must name exact files, scope, blocked actions, validation commands, rollback or fallback expectations, and human approval gates.

If copy/docs crosses into behavior, access, data, automation, or external effects, this route must stop and point to a separate proposal.