EternaVault

An AI-assisted digital inheritance vault that encrypts data during life and releases it only to verified heirs through on-chain identity and validator approval.

Comment

Inspiration

As more of our lives move online, critical personal data - documents, memories, credentials - often becomes permanently inaccessible after death.

Existing solutions rely on passwords, centralized platforms, or legal processes that are slow, fragile, and inconsistent across regions. Even when access is granted, there is no technical guarantee that the right person receives the data at the right time.

We wanted to build a system where digital inheritance is:

  • Private by default
  • Cryptographically enforced
  • Verifiable
  • Human-centered

EternaVault was inspired by the need for a secure, automated, and trustworthy way to pass on digital identity and personal data without relying on platforms or manual judgment.

What it does

EternaVault is a decentralized digital inheritance system that enables users to securely preserve and transfer encrypted digital assets.

It allows users to:

  • Encrypt files client-side so no server can ever read them
  • Assign verified heirs ahead of time
  • Require validator-confirmed events before access is unlocked
  • Use AI to generate meaningful summaries of inherited content after decryption

The system ensures:

  • No plaintext exposure
  • No impersonation or forged access
  • No centralized override
  • Fully auditable access control

If authorization conditions are not met, access is cryptographically impossible.

How we built it

Frontend

  • React + Vite
  • Role-based dashboards (Owner, Heir, Validator)
  • Local encryption and decryption in the browser

Security

  • AES-GCM client-side encryption
  • Keys never leave the user’s device

Backend

  • Node.js + Express
  • Handles encrypted uploads and metadata only
  • Zero plaintext processing

Storage

  • Supabase Storage for encrypted blobs
  • Supabase Postgres for metadata (IV, salt, state)

Blockchain

  • Solidity smart contract deployed on QIE Mainnet
  • Governs identity, validator approval, and access checks

AI Layer

  • Uses OpenRouter to generate AI-based legacy summaries
  • AI operates only on user-decrypted data, never raw ciphertext

Architecture Client encrypts → backend stores ciphertext → smart contract enforces access → heir decrypts locally → AI processes decrypted content (optional)

AI Use Case

AI is used to transform inherited data into meaningful, human-readable summaries once access is authorized.

This enables:

  • Contextual understanding of documents or memories
  • Emotional and narrative preservation
  • Reduced cognitive load for heirs handling large datasets

AI is intentionally isolated from encryption and access control to preserve privacy and trust.

Challenges we ran into

  • Designing an AI-enhanced system without compromising privacy
  • Preventing any AI interaction before cryptographic authorization
  • Ensuring the backend has zero trust and zero decryption capability
  • Balancing strong security guarantees with a simple user experience
  • Deploying and testing real mainnet smart contract logic

Accomplishments we’re proud of

  • Built a fully functional end-to-end prototype
  • Achieved complete client-side encryption with zero plaintext exposure
  • Deployed and interacted with a live mainnet smart contract
  • Integrated AI without violating privacy or trust boundaries
  • Delivered a scalable, real-world applicable solution

Impact & Usefulness

EternaVault addresses a growing real-world problem as digital identity and data ownership expand.

Potential use cases include:

  • Personal digital legacy management
  • Secure document inheritance
  • Identity-controlled data escrow
  • Long-term archival of sensitive data

The system is designed to scale across regions without relying on centralized authorities.

What we learned

  • AI systems must respect cryptographic boundaries to maintain trust
  • Privacy-by-design simplifies long-term security
  • Clear architecture improves both UX and auditability
  • Real-world impact comes from restraint, not feature overload

Future Scope

  • Multi-validator quorum approval
  • Zero-knowledge identity proofs
  • Expanded AI-assisted legacy tools
  • Multi-heir and conditional inheritance logic
  • Integration with formal DID standards

Built With

Share this project:

Updates