![D[AI]NGEN](/brand/logo-dark.png){kind=logo}
SECURITY
Security should be verifiable.
D[AI]NGEN is built around the principle that AI authorization should be evaluated independently from user authorization. Our public security communication follows the same rule: we distinguish verified product behavior from architecture concepts, roadmap direction, and future capabilities.
CORE PRINCIPLE
User access ≠ AI access
Traditional enterprise access decisions answer whether a person or application may access a resource.
AI changes the context. A legitimate employee may use an AI application that has different trust, processing, destination, or data-handling characteristics.
D[AI]NGEN treats the AI system as a separate access context rather than assuming it should inherit the user’s authorization.
The employee’s normal Windows access remains intact. Each AI application receives a separate and more restricted scope.
USER DATA SPACE
COMPLEMENTARY CONTROL LAYER
D[AI]NGEN does not replace IAM, DLP, EDR, CASB, or SSE.
IAM answers who the user is and what the user may access. D[AI]NGEN answers what this AI application or model may use or access.
D[AI]NGEN is designed to complement established identity, data-security, endpoint, and network controls by making the AI-specific authorization boundary explicit.
ADMINISTRATIVE AUTHORITY
Discovery does not authorize access.
Discovery is designed to understand the environment and recommend useful configuration. It does not replace the administrator decision or the policy application step.
The administrator defines or approves which AI application or model may be available, which Workspace or resource is shared, whether a recommendation is accepted, and where applicable how long access remains active.
D[AI]NGEN discovers, recommends, applies, blocks, records, and proves. The administrator remains the authority for access decisions.
DECISION ≠ ENFORCEMENT
An approval is not proof of enforcement.
A management decision and an enforced endpoint state are not treated as the same thing.
The D[AI]NGEN product model requires the endpoint application of the intended access state before that state is considered effective.
FAIL-CLOSED PRODUCT PRINCIPLE
Explicit protection is not silently overridden.
The D[AI]NGEN product model is designed so a newly requested AI share does not silently override an explicitly protected scope.
Conflicting access should fail closed rather than expanding AI access implicitly.
CLOUD / ENDPOINT BOUNDARY
Keep filesystem reality at the endpoint.
D[AI]NGEN is designed so local filesystem resolution and raw content analysis remain on the endpoint where possible.
Central administration expresses intent and receives the metadata and product evidence required for review. This keeps the control model aligned with the device where the data actually exists.
This is a public product concept, not disclosure of internal implementation mechanics.
PUBLIC MODEL
The boundary model
This is the public conceptual security model. It describes the intended security boundary and decision flow. It does not imply that every input or enforcement mechanism shown is currently a verified shipping capability.
ALLOW
Conceptually permit the interaction.
REDACT
Conceptually remove sensitive content before proceeding.
DENY
Conceptually prevent the interaction.
FROM CONCEPT TO CLAIM
We distinguish what is designed, what is implemented, and what is validated.
D[AI]NGEN intentionally separates product direction from verified product guarantees.
A D[AI]NGEN public claim is intentionally narrower than internal product development. This keeps security communication precise.
CONCEPTUAL
Security model
IMPLEMENTED
Product behavior exists
VALIDATED
Tested within defined scope
PUBLIC CLAIM
Approved for communication
WHY THIS MATTERS
Product evidence is not a security guarantee.
An architecture concept may explain where the product is going. An implementation may demonstrate that work exists. A UI may show product state. A Preview demo may show current behavior.
None of those automatically becomes a public security guarantee. A capability should become a public shipping claim only after its behavior has been validated and approved.
THREE DIFFERENT THINGS
Interface, model, and verified behavior should not be conflated.
Product interface
What the user sees.
Security model
How D[AI]NGEN describes the intended boundary.
Verified security behavior
Behavior that has been validated and approved for public communication.
PRODUCT TRUTH V2
A validated behavior is narrower than a capability category.
A broad capability category can remain UNKNOWN even when a narrower behavior has been validated in a specific product flow.
This second layer lets D[AI]NGEN communicate real product progress without converting a bounded demo or test result into a generalized security guarantee.
BROAD CLAIM
Capability category
A general capability such as policy enforcement or transmission prevention remains unpublished until the category itself is verified.
VALIDATED · LIMITED SCOPE
Validated behavior
A specific behavior can be described when evidence supports it and the public wording includes its boundary.
AVAILABLE OR PREVIEW
Product surface
Availability describes the product surface. It does not make every capability associated with that surface verified.
REDACTION SCOPE
Supported text redaction is validated; broad document redaction is later.
Current validated behavior is limited to supported text findings that can produce a redacted copy in scoped evaluated flows.
Partial-document sharing and broader document-format redaction, including PDFs, remain later product direction and are not presented as current broad capability.
VALIDATED · LIMITED SCOPE
Current
Supported text findings can produce a redacted copy in evaluated flows.
ROADMAP
Later
Broader document and PDF redaction remains future product direction.
PRODUCT TRUTH
Current public status
AVAILABLE NOW
Browser Extension
Current product surface, available through private access.
AVAILABLE NOW · PACKAGED RELEASE
Desktop
Technical maturity: Preview. Packaged release launched: October 21, 2026.
VERIFICATION REQUIRED
Conceptual capabilities
Published as shipping features only after verification.
PUBLIC CLAIMS
What we do not claim
D[AI]NGEN does not publish certifications, penetration-test results, encryption guarantees, isolation guarantees, or compliance claims unless they have been completed and verified.
Security documentation will evolve with the validated product.
RESPONSIBLE DISCLOSURE
Report a security issue
If you believe you have identified a security issue involving D[AI]NGEN or this website, contact: contact@daingen.app
Please include enough information to help us understand and reproduce the issue. Do not include unnecessary sensitive data.
TRUST DOCUMENTATION
Trust material grows with the validated product.
As D[AI]NGEN matures, public trust documentation may expand to cover validated architecture, data handling, deployment guidance, supported environments, and other security-relevant information.
Only verified material should be published.