Changelog - The Journey

Sep 28, 2025

v0.0.9

Repository polish and final preparations.

✨ Final Polish

Clean, simple, and submission-ready.

This release focuses on repository elegance:

Pristine folder structure
Clear navigation paths
No unnecessary files
Perfect for ETH Global review

📁 Repository Structure

speaky/
├── README.md     # Complete project overview with demo video
├── .gitignore    # Essential ignore rules
├── web/          # Voice blockchain interface
├── deck/         # Pitch presentation  
└── changelog/    # Development history

Each application is self-contained with its own documentation.

🎯 Ready for Judging

Live Demo: speaky.wtf
Presentation: deck.speaky.wtf
Progress: changelog.speaky.wtf

Simple. Clean. Revolutionary.

Sep 28, 2025

v0.0.8

ETH Global Hackathon Submission - Voice-Controlled Blockchain Interface

🎉 Submission Ready

Speaky is complete and ready for ETH Global judging!

We're pioneering the future of Web3 interaction through natural voice commands. No more complex interfaces - just speak to the blockchain.

🚀 Live Deployments

All systems operational and demo-ready:

Main App: speaky.wtf - Voice-controlled Ethereum wallet
Pitch Deck: deck.speaky.wtf - Project presentation
Changelog: changelog.speaky.wtf - Development journey

✨ Core Features

Voice-Controlled Wallet Operations

Natural language commands for blockchain interaction
Real-time balance checking with voice
Secure transaction sending through speech
MetaMask integration with voice guidance

Technical Innovation

OpenAI Realtime API for instant voice processing
WebRTC for low-latency audio streaming
Custom tool architecture for blockchain operations
Edge-optimized deployment on Vercel

🏆 Bounty Alignment

ENS Integration Track

Voice commands for human-readable blockchain interaction
Natural language processing for Web3 operations
Making crypto accessible through conversation

Main Track Submission

Revolutionary UX for blockchain interaction
Production-ready implementation
Scalable architecture for future expansion

📚 Documentation

This release includes comprehensive documentation:

Complete technical architecture
Voice command reference
Developer setup guides
Contributing guidelines

🔮 Vision

Speaky is building the voice interface for Web3. Our roadmap includes:

3D holographic assistants
AR experiences bringing blockchain into the real world
Universal voice interface for all blockchain operations

🎯 Try It Now

Visit speaky.wtf
Connect your MetaMask wallet
Click "Start Voice Chat"
Say "What's my balance?" or "Send 0.01 ETH"

Experience the future of blockchain interaction today!

Built with passion for ETH Global Hackathon Making blockchain as natural as conversation

Sep 28, 2025

v0.0.7

Voice-controlled Ethereum wallet assistant - Phase 2 (90% complete).

🎙️ Major Blockchain Integration

Speaky now speaks blockchain! Voice commands can now control your Ethereum wallet, check balances, and send transactions - all through natural conversation.

🚀 What's New

Voice-Controlled Wallet Operations

"What's my balance?" - Check ETH balance instantly
"Send 0.01 ETH" - Send transactions with voice
"Is my wallet connected?" - Connection status check
Natural language understanding for all commands

MetaMask Integration

One-click wallet connection
Secure transaction approval flow
Real-time balance updates
Automatic reconnection support

🛠 Technical Implementation

Blockchain Stack

Ethers.js v6 for Ethereum interactions
Gas estimation with 10% buffer for reliability
Transaction confirmation and receipt tracking
WebRTC-to-blockchain bridge for voice control

AI Assistant Features

Custom "Speaky" persona with blockchain knowledge
Context-aware responses about wallet state
Helpful guidance for crypto beginners
Error handling with clear explanations

Available Tools

get_wallet_balance: Real-time ETH balance
check_wallet_connection: Wallet status verification
send_ethereum: ETH transfers (test address)
show_toast: Visual notifications
estimate_transaction_gas: Coming soon

📈 Development Progress

Phase 2: Wallet Interaction (90% Complete)

✅ MetaMask connection
✅ Balance checking
✅ Transaction sending
✅ Voice command processing
⏳ ENS domain resolution (coming in v0.0.8)

Once ENS is added, Phase 2 will be complete!

🔒 Security

MetaMask approval for all transactions
Test address: 0x742d35Cc6634C0532925a3b844Bc9e7595f0bEb0
No private key exposure
Clear confirmation prompts

🎯 Voice Examples

"Check my balance"
"How much ETH do I have?"
"Send 0.1 ETH"
"Am I connected?"

🔮 Roadmap

Phase 1: ✅ Basic voice chat (Complete)
Phase 2: 🔄 Wallet interaction (90% - missing ENS)
Phase 3: 🔜 3D character with voice
Phase 4: 🔜 WebAR - bring it out of the browser

Next Up (v0.0.8)

ENS resolution to complete Phase 2 - send to "vitalik.eth"!

Demo: https://speaky.wtf

Built for ETH Global Hackathon - Making blockchain as natural as conversation!

Sep 28, 2025

v0.0.6

Project presentation and documentation enhancements.

📊 Presentation Deck

Project Pitch Deck

Interactive presentation explaining Speaky's vision and value proposition
Published at deck.speaky.wtf for easy stakeholder access
Structured narrative covering problem, solution, market, and roadmap

Key Sections

Problem/Solution: Voice-controlled blockchain interaction
Technology Stack: Integration architecture and technical implementation
Market Analysis: Target audience and growth potential
Business Model: Revenue streams and monetization strategy
Product Demo: Live demonstration capabilities
Team & Vision: Project leadership and future roadmap

🎨 Additional Improvements

Enhanced favicon implementation across all applications
Wallet integration improvements in the main web app
UI refinements and bug fixes

Technical Details

Built with presentation framework for smooth navigation
Responsive design for all device sizes
Optimized for Vercel deployment

View the deck at: https://deck.speaky.wtf

Sep 27, 2025

v0.0.5

Nuxt 4 migration with unified architecture and toast notifications.

🚀 Architecture Evolution

From React to Nuxt 4

Complete migration from React + Express to Nuxt 4
Unified full-stack application under single framework
Simplified deployment and development workflow
Previous React implementation preserved in trials/

🎨 UI Enhancements

Nuxt UI Integration

Modern component library replacing React components
Professional design with consistent styling
Dark mode support out of the box
Responsive layout with semantic HTML structure

Interface Components

Clean VoiceChat component with session management
Status indicators for connection state
Text input alongside voice interaction
Navigation with links to deck and changelog

🔧 Technical Implementation

Composables Architecture

useRealtimeChat: WebRTC session management
useRealtimeTools: Function calling implementation
Clean separation of concerns
TypeScript for type safety

Server API Routes

/api/session: SDP exchange for WebRTC
/api/token: Ephemeral token generation
Runtime config for API key management
Error handling with proper HTTP status codes

🎯 Function Calling

Toast Notifications

ChatGPT can trigger toast notifications
Support for success, error, warning, info types
Custom titles, descriptions, and icons
Bidirectional communication via data channel

📦 Developer Experience

Simplified Stack

Single pnpm dev for full-stack development
Nuxt DevTools for debugging
Hot module replacement
TypeScript throughout

🔄 Migration Path

The journey from React to Nuxt:

Started with OpenAI's React template
Identified need for simpler architecture
Migrated to Nuxt 4 for unified experience
Added Nuxt UI for professional components
Implemented toast notifications as proof of concept

Next step: Integrate ethers.js for Web3 capabilities

Sep 27, 2025

v0.0.4

Speech-to-speech AI chat implementation using OpenAI Realtime Console.

🎤 Speech-to-Speech AI Chat

Working Implementation

Successfully integrated OpenAI's official Realtime Console template
Speech-to-speech conversation with AI now functional
Based on https://github.com/openai/openai-realtime-console

🛠 Journey to Success

Initial Challenges

Multiple implementation attempts with various approaches
Troubleshooting WebRTC vs WebSocket configurations
API model naming and endpoint issues
Authentication and token generation complexities

Final Solution

Used OpenAI's official example as foundation
WebRTC implementation for optimal audio performance
Proper ephemeral token authentication
React + Vite + Express.js stack

📁 Project Structure

Clean Organization

Working implementation in /web/ directory
Trial attempts moved to /trials/ (gitignored)
Official OpenAI template successfully integrated
Ready for customization and feature development

🚀 Development Setup

cd web
npm install
cp .env.example .env
# Add your OpenAI API key to .env
npm run dev

🎯 Next Steps

With the foundation working, we can now:

Customize the React interface to our needs
Remove unnecessary template bloat
Add custom features and styling
Build the voice chat experience we envision

The speech-to-speech journey begins here! 🗣️✨

Sep 26, 2025

v0.0.3

Vercel deployment fixes and branding improvements.

🔧 Bug Fixes

Vercel Deployment Build Metadata Issue

Fixed 404 errors for /_nuxt/builds/meta/*.json after initial page load
Root cause: Build ID mismatch between client and server
Solution: Added nitro: { preset: "vercel" } for proper Vercel configuration
Disabled experimental app manifest feature to prevent build conflicts
Cleared all build artifacts and rebuilt with fresh configuration

🎨 UI Improvements

Logo Layout Enhancement

Fixed "Speaky" text positioning from below to beside the SVG icon
Implemented horizontal flex layout with proper alignment and spacing

Custom Branding

Updated page title to "Changelog - The Journey"
Enhanced description: "Track the evolution and progress of our projects through detailed release notes and development milestones"
Streamlined navigation by removing Documentation link
Applied branding changes to both meta tags and UI content