VoiceLegacy

Preserve the voice, words, and phrases that make communication feel personal before speech loss occurs.

Every year, millions of people lose the ability to speak — due to ALS, stroke, throat cancer, Parkinson's, or progressive neurological conditions. VoiceLegacy lets people preserve their natural voice and build a personal communication toolkit — so they can always sound like themselves when it matters most.

---

What It Does

1. Voice Cloning — Read 8 short phrases into your microphone. ElevenLabs creates a private clone of your voice in seconds. From that point on, any text can be spoken back in your voice.

2. Phrase Bank — Build a personal library of the things you actually say, organized by category: Family, Daily Needs, Comfort, Humor, Emergency, Personal. Every phrase is playable in your cloned voice with one tap.

3. AI-Powered Communication — Gemini suggests phrases, rewrites messages to sound warmer or shorter, and restyles text to match your saved communication tone. You type what you want to say. The AI shapes it. Your preserved voice speaks it.

4. Privacy-First Design — Explicit consent before anything is recorded. All data is private, never shared, and deletable at any time.

---

Tech Stack

Layer	Technology
Framework	Next.js 16 (App Router, Turbopack)
Styling	Tailwind CSS v4
Authentication	Clerk (@clerk/nextjs v7)
Database	MongoDB Atlas
Voice Cloning + TTS	ElevenLabs API
AI Reasoning	Gemini API (Flash 2.0)
Audio Recording	Browser MediaRecorder API
Deployment	Vercel

---

Getting Started

Prerequisites

• Node.js 20+
• npm 9+
• API keys for: Clerk, MongoDB Atlas, ElevenLabs, Gemini (optional)

Setup

# Clone the repo
git clone https://github.com/ntoptchi/VoiceLegacy.git
cd VoiceLegacy/frontend

# Install dependencies
npm install

# Copy the env template and fill in your keys
cp .env.local.example .env.local

# Start the dev server
npm run dev

Open http://localhost:3000 in your browser.

Environment Variables

Copy frontend/.env.local.example to frontend/.env.local and fill in the values:

Variable	Required	Source
NEXT_PUBLIC_CLERK_PUBLISHABLE_KEY	Yes	dashboard.clerk.com
CLERK_SECRET_KEY	Yes	dashboard.clerk.com
MONGODB_URI	Yes	cloud.mongodb.com
ELEVENLABS_API_KEY	Yes	elevenlabs.io
GEMINI_API_KEY	Optional	aistudio.google.com

If you don't have a key yet, set the matching MOCK_* flag to true:

MOCK_DB=true            # In-memory store, no MongoDB needed
MOCK_VOICE_API=true     # Stub ElevenLabs responses
MOCK_GEMINI_API=true    # Stub Gemini responses

---

User Flow

Landing Page (/)          → No auth required
    ↓
Record Voice (/record)    → No auth — record 8 phrases, create voice clone
    ↓
Preview (/preview)        → No auth — hear your clone speak for the first time
    ↓
Sign Up (Clerk)           → Create account after hearing the emotional hook
    ↓
Consent (/consent)        → Privacy agreement + communication style
    ↓
Phrase Bank (/phrases)    → Build your personal phrase library
    ↓
Speak For Me (/speak)     → AI rewrites + TTS playback in your voice
    ↓
Dashboard (/dashboard)    → Profile, export, delete data

---

Project Structure

VoiceLegacy/
├── frontend/
│   ├── app/                    # Next.js pages and API routes
│   │   ├── api/                # Backend API routes
│   │   │   ├── user/           # User CRUD (Clerk auth)
│   │   │   ├── voice/          # Upload, claim, delete voice
│   │   │   ├── phrases/        # Phrase CRUD
│   │   │   ├── gemini/         # AI suggest + rewrite
│   │   │   └── speak/          # Text-to-speech
│   │   ├── record/             # Voice recording page
│   │   ├── preview/            # Clone preview page
│   │   ├── consent/            # Privacy consent page
│   │   ├── phrases/            # Phrase bank page
│   │   ├── speak/              # Speak For Me page
│   │   └── dashboard/          # User dashboard
│   ├── src/
│   │   ├── components/         # Shared UI components
│   │   └── lib/                # Backend helpers, DB, API clients
│   ├── public/                 # Static assets
│   ├── scripts/                # Seed + index scripts
│   └── .env.local.example      # Environment variable template
├── voicelegacy-plan.md         # Full project plan
├── PROGRESS.md                 # Build progress tracker
├── PITCH.md                    # Pitch guide for judges
└── SCRIPT.md                   # Demo video script

---

Scripts

# Start development server
npm run dev

# Build for production
npm run build

# Seed demo data (requires MONGODB_URI in .env.local)
npx tsx scripts/seed-demo.ts

# Create database indexes
npx tsx scripts/ensure-indexes.ts

---

Deployment

The app deploys to Vercel:

1. Import the GitHub repo on vercel.com
2. Set Root Directory to frontend
3. Import environment variables from .env.local
4. In MongoDB Atlas, add 0.0.0.0/0 to the IP Access List for serverless
5. In Clerk, add the Vercel domain to allowed origins
6. Deploy

---

Team

Built at Hackabull by:

• Nicholas Toptchi
• Jackson Bopp
• Amrit Selva Ganesh
• Connor Kouznetsov

---

License

This project was built for a hackathon. All rights reserved.