indentia.ai

Security & trust

Cryptographic where it counts. Auditable everywhere.

Indentia is built for environments where "we trust the network" is not an acceptable answer. Every request is evaluated at runtime; every artefact carries a cryptographically-bound label; every key lives in hardware. Defence-grade controls, available to anyone who needs them, not just to ministries.

Indentia is data centric security: the perimeter is not the point of defence, the data itself is.

Discuss your threat model Back to the platform

Zero-trust access chain: every request evaluated at runtime on identity, intent, context and resource before touching data

Capabilities

Six controls, designed as one system.

Zero Trust by default

Every request is evaluated at runtime (identity, intent, context and resource) before it touches anything. No implicit trust based on network position, VPN, namespace or service-account. A request from inside the cluster gets the same scrutiny as one from outside.

Semantic ABAC + RBAC

Attribute-Based Access Control over the ontology, layered on classical role-based access. Permissions are evaluated against entity attributes (clearance, releasability, domain, sensitivity) and dynamic context, not just on group membership. Automated, expressed in policy-as-code.

Cryptographic classification labels

STANAG 4774 confidentiality labels and STANAG 4778 metadata bindings: every artefact carries a label, cryptographically bound to its content via XMLDSIG. Downgrade-attacks are detectable; releasability checks happen on the label, not on a separate database lookup.

Provenance & chain-of-custody

Every fact in the knowledge graph carries its source, ingestion time, transformations and the responsible identity. Tampering is detectable; "who told us, when, and how was it transformed" is a SPARQL query, not a forensic project.

PKI + mTLS everywhere

Mutual TLS between every internal service, with certificates issued from an internal PKI you control. Service-to-service trust is cryptographic, not network-topological. Rotation is automated; revocation is observable.

HSM-backed keys (EAL4+ / FIPS 140-3 L3)

Signing keys, sealing keys and at-rest encryption keys live in a hardware security module that meets EAL4+ Common Criteria or FIPS 140-3 Level 3, depending on your environment. The platform never sees raw key material. Operations happen inside the HSM.

Where it fits

Built for the toughest environments. Useful everywhere else.

Defence & government

Compartmented data with STANAG labels, releasability enforced at the platform layer, HSM-protected signing for orders and intelligence products.

Banking & insurance

Customer-data segmentation per business unit and jurisdiction, audit-grade chain-of-custody for every model decision, regulator-ready evidence on demand.

Critical infrastructure

NIS2-compliant zero-trust between IT and OT, signed firmware and configuration for field devices, immutable telemetry chain.

Proof, not promises

Four tests you can run in a day.

Do not take our word for it. These are practical tests you can execute yourself, without any customer data, and without us in the room.

1 · Air-gapped installation

A full environment without an internet connection, including the management and update path. Your test: pull the plug and keep working.

2 · Meaningful export

Data, ontologies, links, identifiers and lineage in open formats. Your test: export it and read it outside our platform.

3 · After-the-fact reconstruction

An answer from weeks ago, traced back to source, definition, version and authorisation. Your test: pick any old answer and request the trail.

4 · Permissions in the chain

The same question yields a different answer for a different role. Your test: log in as two roles and compare.

Sovereign by construction

Your keys. Your code. Your hardware.

PKI, HSM, signing keys, sealing keys: all under your control, inside your perimeter. Indentia can run with zero outbound connectivity. No SaaS dependency in the trust chain; no foreign jurisdiction in the audit trail.

Request a security deep-dive