Aqua Protocol · Version 4

In development

AI is moving fast.
Security can't keep up.

  • Who created this AI output?
  • Can I trust this data source?
  • Was this model tampered with?
  • Is this identity verified?
  • Who authorized this access?

Aqua V4 is the trust infrastructure for the AI era: identity, access control, and provenance. Cryptographic proof that the data you use, produce, and share is real.

The problem

The AI trust crisis

AI is generating content, making decisions, and accessing systems at unprecedented scale. But there is no infrastructure to verify any of it.

Industry signal · Lex Fridman
“Very soon, if not already, security will become THE bottleneck for effectiveness and usefulness of AI agents as a whole, since intelligence is still rapidly scaling and is no longer an obvious bottleneck for many use-cases.”
Lex identifies three pillars of AI agent power:
1.Model intelligenceScaling
2.Data accessUnsecured
3.Agent autonomyUnsecured
Source
Regulatory context

It is not just the market. Regulators on both sides of the Atlantic are converging on the same conclusion: AI systems need verifiable provenance, identity, and audit infrastructure. This is becoming law.

US: NTIA RFC 2023-07776

The National Telecommunications and Information Administration called for public input on AI accountability, receiving 1,400+ comments. The resulting 2024 report explicitly recommends investment in:

Rec. 5: “Provenance technologies” to assess AI training data and usage
Rec. 2: Standard disclosures for AI architecture, data, and performance
Rec. 6: Independent evaluations of high-risk AI systems
EU: AI Act (2024/1689)

The world's first comprehensive AI regulation. Entered into force August 2024, with obligations phasing in through 2026. Non-compliance carries severe penalties.

Art. 10: Data governance: provenance documentation, source traceability
Art. 12: Automatic logging for traceability and post-market monitoring
Art. 13: Transparency obligations for AI system providers
Full text Penalty: up to 7% of global turnover

Identity crisis

AI agents act on behalf of users, but there is no way to verify who they represent. Spoofed identities, unauthorized delegation, zero accountability.

> identity.verify() → UNDEFINED

Access chaos

AI models consume data from everywhere. No granular access control. No audit trail. No way to know what data was used, when, or by whom.

> access.log() → NULL

Provenance gap

Where did this data come from? Was it tampered with? Is this output real or generated? Without provenance, AI outputs are just noise.

> data.origin() → UNKNOWN

The solution

Aqua V4: trust infrastructure

A cryptographic protocol that gives AI systems what they are missing: verifiable identity, granular access control, and tamper-proof provenance.

Identity

Self-sovereign identity with cryptographic attestation. Every agent, every model, every user, verified and accountable.

  • Peer-to-peer attestation
  • Challenge-based verification
  • No central authority needed

Access control

Granular, cryptographic access policies. Control exactly who and what can read, write, or delegate across your data.

  • Revision-level permissions
  • Auditable access trails
  • Zero-trust architecture

Provenance

Every piece of data carries its full history. Tamper-proof, portable, and independently verifiable, with no blockchain required.

  • Cryptographic hash chains
  • Portable verification
  • Instant tamper detection

How it works

Pillars of verifiable trust

Aqua combines cryptographic hash chains, modular signatures, optional timestamping, and decentralized verification to create trust without central authorities.

Portable hash-chains

Aqua creates AquaTrees, portable data structures that record a complete history of revisions with cryptographic precision. Every change is chained, every version is preserved.

> Each revision linked by hash to its predecessor

Modular signatures

Aqua supports multiple signature schemes: Ethereum wallets, DIDs, X.509 certificates, and more. The signature layer is modular by design: plug in the cryptographic paradigm that fits your security requirements without changing the protocol.

> One protocol, any signature scheme

Immutable timestamps

Anchor your data to tamper-proof timelines via the Ethereum blockchain, qualified TSA services (eIDAS-compliant), or both. Blockchain is optional; Aqua integrates with existing institutional infrastructure and decentralized networks alike.

> Blockchain or TSA: choose your trust anchor

Flexible trust

Aqua bridges institutional and decentralized worlds. Use enterprise registries for regulatory compliance or community-driven decentralized registries for open verification, or combine both. One protocol, every trust model.

> Institutional or decentralized: works with both

Open protocol

The Kubernetes of AI verification

Aqua V4 is fully open source. Trust infrastructure cannot depend on a single vendor; it must be auditable, forkable, and verifiable by anyone. Free to build on. Enterprises deploy with production-grade infrastructure, compliance, and SLAs.

Open spec

Full protocol specification, verification trees, and SDK, all publicly auditable.

Open standard

Interoperable by design. Every implementation strengthens the verification network.

Enterprise ready

Build free. Deploy with production-grade infrastructure, SLAs, and compliance.

Who it's for

One protocol. Every stakeholder.

Operate

Operational trust for AI agents. Identity proves who an agent is; authorization defines what it may do. Agents cannot exceed their boundaries, and every action leaves verifiable proof of what actually occurred.

Explore operational trust

Invest

Trust infrastructure for AI: regulatory tailwinds, enterprise monetization, and a massive unmet market. The security layer AI has been missing.

Book a call

License & integrate

Enterprise-grade, compliance-ready deployment. Custom integration, licensing, and production infrastructure for your organization.

Request a demo

Build

Open protocol, open SDK. Ship verified AI applications with cryptographic trust baked in from day one.

View on GitHub

See it in action

From vulnerable to verified

aqua-v4 · trust-pipeline
$ aqua init --protocol v4
Initializing Aqua trust chain...
✓ Genesis revision created
$ aqua sign --identity did:aqua:0x7f3a...
Binding cryptographic identity...
✓ Identity attested and sealed
$ aqua verify --file ai-output.json
Verifying provenance chain...
Checking 14 revisions across 3 witnesses...
✓ INTEGRITY VERIFIED: all hashes match
✓ IDENTITY CONFIRMED: signer verified
✓ PROVENANCE VALID: complete chain from origin
$

The difference

Without vs. with Aqua

Without Aqua

  • AI outputs with no verifiable origin
  • Identities that can be spoofed trivially
  • No audit trail for data access
  • Tampering is undetectable
  • Trust based on faith, not proof
  • Central authorities as single points of failure

With Aqua V4

  • Every output carries cryptographic provenance
  • Self-sovereign identity with peer attestation
  • Complete, immutable access audit trail
  • Instant tamper detection via hash verification
  • Trust based on mathematics, not authority
  • Decentralized, with no single point of failure

Work with us on Aqua V4

Version 4 is in active development as a proposed open standard. If you are evaluating it, integrating it, or would like a technical walkthrough, we are happy to talk.

AquaAqua Protocol V4

© 2026 inblock.io assets GmbH. All rights reserved.