CaniTrust by CAIOS

A privacy-first verifiable identity layer for pets, owners and AI-powered care.

Comment

Inspiration

A companion animal can have a microchip, a local-authority registration, veterinary records, insurance records, an owner profile and multiple digital identities.

Yet these records often live in separate systems.

The problem becomes even more important as AI agents begin helping owners manage feeding, health observations, daily care and other decisions.

Before an AI system asks:

“What happened to this pet?”

there is an even more fundamental question:

“Can we verify who this pet is?”

That question inspired CaniTrust.

What it does

CaniTrust is a privacy-first Verifiable Companion Identity layer.

Our goal is simple:

One Pet. One Verifiable Identity. Trusted Everywhere.

CaniTrust links three important identity relationships:

  1. Physical Pet Identity
    Such as a microchip or registered pet identifier.

  2. Authority / Registration Credential
    A credential issued by, or derived from, an authorised registry or local authority.

  3. Owner Relationship Credential
    A verifiable relationship between the pet and its authorised owner.

Together, these form a persistent:

Pet Trust ID

The Pet Trust ID can later be verified by authorised parties such as veterinary providers, shelters, insurers, boarding providers, pet-care platforms and AI agents.

Privacy First: What We Do NOT Put On-Chain

CaniTrust does not propose putting a pet owner's identity card number, address, phone number, full pet medical record or other sensitive personal information on a blockchain.

Instead, we use a Selective Proof / Privacy-First architecture.

Sensitive information remains off-chain with authorised systems.

The verification layer stores only the minimum cryptographic information necessary to establish properties such as:

  • credential integrity;
  • issuer provenance;
  • credential status;
  • timestamps; and
  • revocation status.

This allows a relying party to verify that a credential is authentic and has not been altered without exposing the underlying private information.

Web3 should verify trust, not expose identity.

How It Works

The proposed CaniTrust identity flow is:

Pet Identifier / Microchip

↓

Authority or Registry Credential

↓

Owner Relationship Credential

↓

CaniTrust Pet Trust ID

↓

Selective Cryptographic Proof

↓

Authorised Verification

A veterinary provider, shelter, insurer, boarding provider or AI system could therefore verify the identity relationship without receiving unnecessary personal information.

Example: Pika

Imagine a dog named Pika.

Pika has a microchip identifier and an owner.

In our Stage 1 concept:

  1. Pika's pet identity is created.
  2. A Mock Authority Issuer issues a registration credential.
  3. An Owner Relationship Credential links Pika to the authorised owner.
  4. CaniTrust creates a selective cryptographic proof.
  5. An authorised relying party verifies:

✓ Pet identity valid
✓ Credential integrity valid
✓ Ownership relationship valid
✓ Credential status valid

The verifier does not need to see the owner's government ID number or Pika's private health records.

Relationship With AI Agents

CaniTrust is designed to become a trust boundary for companion-care AI.

An AI agent should not only ask:

“What data do I have?”

It should also be able to ask:

“Can I trust the identity and provenance behind this data?”

This creates a future architecture:

Identity → Memory → Context → Intelligence → Action

CaniTrust focuses on the first layer:

Identity.

How We Built the Stage 1 Concept

For NTU InnovateX Stage 1, we designed CaniTrust around:

  • privacy-first identity architecture;
  • W3C-style Verifiable Credential concepts;
  • decentralised identity concepts;
  • deterministic cryptographic hashing;
  • selective proof;
  • credential status and revocation;
  • off-chain private credential storage;
  • EVM-compatible testnet anchoring as a proposed verification mechanism; and
  • a Mock Authority Issuer for prototype development.

The architecture deliberately separates private data storage from public verification.

Challenges We Faced

Our biggest design challenge was deciding:

What actually belongs on Web3?

It would have been easy to place pet records, care events or owner information on-chain.

We rejected that approach.

Companion identity involves real people, sensitive records and potentially long-lived personal information.

Therefore our design principle became:

Put trust proofs on the verification layer — not private lives.

A second challenge was avoiding unnecessary blockchain complexity.

CaniTrust does not require every feeding record, daily activity or AI interaction to become a blockchain transaction.

Web3 is used selectively where verification creates real value.

What We Learned

We started by exploring how Web3 could verify pet-care records.

That led us to a more fundamental insight:

Before an intelligent system can build trusted memory around a pet, it needs a trusted answer to “Who is this pet?”

This shifted CaniTrust from a care-event verification concept into a Verifiable Companion Identity infrastructure.

What's Next

The next prototype milestones are:

  1. Pet Trust ID creation;
  2. Mock Authority Credential issuance;
  3. Owner Relationship Credential issuance;
  4. selective cryptographic proof generation;
  5. QR-based credential verification;
  6. credential revocation;
  7. EVM testnet proof anchoring; and
  8. a narrow integration boundary for authorised AI agents.

Longer term, CaniTrust could support interoperability between:

  • pet owners;
  • local authorities and registries;
  • veterinarians;
  • shelters and adoption organisations;
  • insurers;
  • pet hotels and boarding providers;
  • airlines and travel services; and
  • AI-powered companion-care platforms.

CaniTrust is not about creating another token.

It is about creating portable trust around the identity of a living companion.

Prior Work Disclosure

CAIOS is a pre-existing companion-animal intelligence platform developed independently before NTU InnovateX 2026.

CAIOS and its existing memory, context, recommendation, nutrition and other proprietary technology remain separate Background IP.

CaniTrust's Verifiable Companion Identity layer is the new InnovateX project scope.

For the hackathon prototype, authority integration is represented by a clearly labelled Mock Authority Issuer.

We do not claim current integration with, endorsement by, or verification from any government authority.

CAIOS remembers the pet.

CaniTrust proves who the pet is.

Built With

  • cryptographic
  • decentralized-identity
  • did
  • evm
  • hashing
  • next.js
  • privacy
  • qr-code
  • selective
  • smart-contracts
  • typescript
  • verifiable-credentials
  • web3
Share this project:

Updates

Submission history