eStory decentralized AI powered application.

What is it?
 eStory is a Web3 AI-powered voice journaling, memory/story and human meaning dApp that transforms personal narratives into sovereign,   
 verifiable memory infrastructure. Users record voice journal entries, which are transcribed via
 ElevenLabs, enhanced and analyzed by Google Gemini AI, encrypted client-side (AES-256-GCM), and optionally
  minted as NFTs on Base.

 How it works:
 Users speak their stories, which get transcribed, AI-analyzed for significance/emotional depth/quality,  
 and stored with client-side encryption in a local vault. The key innovation is a privacy-preserving    
 dual-write architecture powered by Chainlink CRE: AI analysis runs inside Chainlink's encrypted DON    
 enclave, producing consensus-validated metrics. Only minimal cryptographic proofs (quality tier, threshold
  boolean, metrics hash, author commitment hash) are written on-chain — full scores and themes stay     
 off-chain in Supabase, visible only to the author. Anyone can verify the analysis is legitimate without  
 seeing its contents.

 What problem it solves:

• Voice first approach to reduce friction in journaling so people can write more and be excited to
•  Journals are deeply personal — users self-censor if analysis results are public. Traditional AI analysis  
   requires trusting the app operator. eStory solves both: Chainlink CRE provides verifiable, tamper-proof AI
    computation while the privacy-preserving design ensures sensitive analysis never hits the blockchain.   
   Authors can prove their stories were fairly analyzed without broadcasting scores to the world.
•  eStory is a convergent solution to three compounding crises in human civilization:

 The Ancient Crisis — Memory Extinction. Throughout history, the vast majority of human experience has   
 simply vanished. Oral traditions died with their keepers. Before recording technology, entire lifetimes of
  personal narrative, the daily emotions, reflections, turning points — were permanently lost. We remember
  less than 1% of our lived experience. Humanity has always lacked infrastructure to capture and preserve  
 the individual story at scale.

 The Present Crisis — The Unexamined Life at Scale. We are living through a civilizational inflection point
  — the AI revolution — comparable to the Renaissance and the Industrial Revolution. Future generations   
 will look back at this era and ask how we navigated it, what we felt, what we feared, what we hoped for.  
 Yet most people are too overwhelmed to document it. Journaling declutters the mind, reduces anxiety, and  
 sharpens thinking — science has proven this repeatedly — but platforms that could enable it harvest our  
 most intimate data with zero transparency. AI systems analyze our behavior, but we can't verify their   
 outputs or trust their operators. There's no ownership of insights derived from your own words, and no   
 privacy guarantee that survives a terms-of-service update. The people living through this historic moment 
 need a sovereign space to process it — and a guarantee that their reflections remain theirs.

 The Emerging Crisis — The Meaning Void. As AI automates work at an accelerating pace, millions will lose  
 not just jobs but their primary source of identity and purpose. When the thing you did for a living no   
 longer needs you, an existential vacuum opens. People will search for what fills that gap — and
 storytelling, journaling, and creative self-expression are among the oldest and most fulfilling answers  
 humanity has ever found. But in a world flooded with AI-generated content, how do you prove your story is 
 authentically yours? Without verifiable provenance, genuine human narrative drowns in synthetic noise. The
  tools people turn to for meaning must also protect the authenticity of what they create.

 eStory addresses all three simultaneously: voice capture preserves narratives that would otherwise be lost
  (ancient), client-side encryption and Chainlink CRE provide verifiable, privacy-preserving AI analysis  
 without trusting the app operator (present), and on-chain cryptographic proofs create tamper-proof     
 provenance for authentic human stories in a world that increasingly struggles to distinguish real from   
 generated (emerging).

• Frontend: Next.js 15, React 19, Tailwind CSS 4, shadcn/ui, RainbowKit
   - Web3: Wagmi 2, Viem, Hardhat, Base Sepolia (ERC20 token, ERC721 story NFTs, PrivateVerifiedMetrics    
   contract)
   - AI: ElevenLabs Scribe (speech-to-text), Google Gemini 2.5 Flash (analysis), Claude SDK (thinking agent) 
   - CRE Workflow: 8-step pipeline — parse → Gemini via ConfidentialHTTPClient → derive privacy fields    
   (keccak256 hashes) → ABI-encode minimal data → generate CRE report → write on-chain → callback full    
   metrics to Supabase
   - Encryption: Client-side AES-256-GCM vault with PIN-derived keys (PBKDF2 → KEK → AES-KW wraps DEK)    
   - Smart Contract: PrivateVerifiedMetrics.sol inherits Chainlink's ReceiverTemplate, validates
   KeystoneForwarder as caller, stores only proofs + hashes