Security
A system this sensitive has to be inspectable.
Ownership records, executed agreements and transaction plans are among the most sensitive data a company holds. This page describes architecture rather than badges, and lists only controls that are implemented.
Tenant isolation
Data is scoped per tenant, with object-level permissions applied to records, search, export and graph traversal alike.
Identity
Multi-factor authentication, with enterprise SSO, SAML and SCIM as product readiness allows.
Evidence integrity
Documents are versioned and content-hashed, so an exported bundle can be checked against what the platform holds.
Material audit events
Ownership, rights and verification changes are appended with actor, capacity and before/after state.
Confidentiality labels
Labels constrain traversal, search, export and AI retrieval — not only which page a person can open.
AI data boundary
Retrieval is permission-scoped. Hidden objects do not leak through an answer, a count or a summary.
The detail
How each control is actually implemented.
Written for the person filling in a security questionnaire, who needs specifics rather than adjectives.
- Tenant isolation and authorization
- Every record belongs to a tenant, and authorization is evaluated per object rather than per page. The same check governs a page view, a search result, a graph traversal, an export and a retrieval by the AI layer.
- Encryption
- Data is encrypted in transit and at rest, including evidence documents in object storage.
- Identity
- Multi-factor authentication, with enterprise SSO, SAML and SCIM provisioning as product readiness allows. Where one of these is not yet available it is described as not yet available, here and in a questionnaire.
- Evidence integrity
- Documents are versioned and content-hashed. An exported evidence pack carries a manifest, so a recipient can verify that what they received matches what the platform holds.
- Material audit events
- Ownership, rights and verification changes are appended — not edited — with actor, the capacity they acted in, before/after state and the integration or request that produced them.
- Integration credentials
- Connector tokens use least-privilege, read-oriented scopes and are rotatable without losing a connection's history. The default product reads repository metadata, not source code.
- Backups and retention
- Backups are taken on a schedule and restores are tested; retention is configurable per tenant, including for evidence, which has a longer natural life than most business records.
- AI data boundary
- Retrieval is permission-scoped to the person asking. Hidden objects do not leak through an answer, a count, a summary or a citation, and generated output carries the same labels as the records behind it.
Responsible disclosure
Found something? Tell us before anyone else.
Report a suspected vulnerability to security@intangiblehq.com. A human reads it. Please give us a reasonable window to fix an issue before publishing it, and we will keep you updated rather than going quiet.
What this page does not claim
There are no compliance logos here. Intangible HQ has not completed a SOC 2 or ISO 27001 audit, and publishing a badge before the certificate exists is the kind of claim an enterprise security review catches in the first ten minutes.
If your procurement process requires a specific attestation, tell us which one and we will tell you honestly where it stands.
Bring your security questionnaire.
We would rather answer it in detail now than discover a blocker after you have built a register.
Evidence-backed relationships Multi-entity by design Built for transfer and diligence