Trust, provenance and transparency infrastructure for AI-generated and digital content.

Aligned with EU AI Act Article 50, DSA, enterprise governance, journalism integrity and evidentiary requirements. Axia provides the cryptographic primitives — identity, integrity, disclosure, continuity, traceability — that platforms, publishers and regulators need to mark, verify and audit digital content.

What Axia is — and what it isn't

Axia is not a detector. Not a classifier. Not a moderation engine. It is the cryptographic infrastructure that publishers and platforms use to mark, disclose and verify content. Disclosures are asserted by the publisher and signed — not detected by us.

Five pillars

We don't verify truth. We verify origin, integrity, disclosure, continuity and traceability.

Identity

Who published it — DNS-verified publisher identity bound to every signature.

Provenance

Where it comes from — upstream references, derivatives, editorial chain.

Integrity

Whether it has been altered — Ed25519 signatures over SHA-256 hashes.

Disclosure

Publisher-attested AI / modification status, cryptographically signed.

Transparency

Public, append-only Merkle log — every signature publicly verifiable.

Use cases

Capability mapping

AI provenance taxonomy

All disclosures are asserted by the publisher and cryptographically signed alongside the content hash.

Human-created

Authored entirely by a human. No generative model involved.

AI-assisted

Human-authored with AI tools (autocomplete, suggestions, editing).

AI-generated

Primarily produced by a generative model. Publisher discloses model and review.

AI-modified

Originally human content modified by AI (translation, enhancement, restoration).

EU AI Act alignment

Article 50 of the EU AI Act requires providers and deployers of generative AI to mark synthetic outputs in a machine-readable way and disclose them to users. Axia provides the cryptographic primitives — content hashing, publisher-attested disclosure, public verification and transparency log inclusion — that operators of AI systems and platforms can use to meet these requirements.

Axia is infrastructure, not legal advice. Compliance with EU AI Act, DSA or any other framework depends on the obligations applicable to each specific operator and use case.

Spain — Ley Orgánica de IA (Anteproyecto 2025)

Spain's upcoming Organic Law on AI introduces fines of up to €35 M or 7% of global annual turnover for non-compliance, an AI supervisory authority (AESIA), a Delegado de IA figure for organisations, and explicit obligations around synthetic content and the protection of minors. Axia is non-authoritative infrastructure that supplies the cryptographic primitives operators need to evidence those obligations.

Non-authoritative mapping. Axia does not classify AI systems and does not replace the Delegado de IA. The competent authority — AESIA in Spain — decides admissibility and compliance.

Content safety policy at the signing layer

Beyond marking and disclosure, Axia enforces hard rules at the signing API. These rules apply to every signature, regardless of plan or integration method.

• Refuse to sign content declared as synthetic sexual material.
• Block signing of ai_generated content without an explicit human_review = true attestation.
• Mandatory content_origin declaration (human | ai_assisted | ai_generated | captured) on every signature.

Axia does not detect synthetic sexual content. It refuses to attest it once declared, and refuses ai_generated signatures without human supervision.

Start signing and disclosing today

Verify a file, attach a publisher-attested AI disclosure and get a public proof URL in seconds.