666FR3K v2.0

🔥 AI-Powered Auto-Improvement System 🔥

A revolutionary auto-improvement loop that monitors agent conversations, extracts insights, and automatically integrates improvements into workflows and codebases. Features continuous speech recognition & synthesis, real-time monitoring, and dynamic audio generation that never repeats.

🚀 NEW in v2.0: Auto-Improvement Loop

Revolutionary Features:

• 🧠 Agent Conversation Monitoring: Real-time monitoring of all agent communications
• 💡 AI-Powered Insight Extraction: Automatically identifies patterns, improvements, and optimizations
• 🔄 Automatic Workflow Integration: Insights are integrated into CLAUDE.md and workflow files
• 💻 Dynamic Codebase Updates: Auto-applies code improvements based on discovered patterns
• 🔊 Evolving Audio: Generates unique audio updates that never repeat (hash-verified)
• 📚 Self-Adapting Knowledge Base: Continuously learns and improves from conversations
• 🌐 Real-Time Web Dashboard: Monitor system status, insights, and metrics live
• ♾️ Continuous Evolution: Runs indefinitely, constantly improving the system

🎯 Core Features

• 🔊 Text-to-Speech: Convert text to natural speech using gTTS (Google Text-to-Speech)
• 🎤 Speech-to-Text: Real-time transcription using Google Speech Recognition
• 🔄 Continuous Loop: Automated TTS→Audio→STT testing pipeline
• 🎧 Auto-Listener: Continuous microphone monitoring with voice command support
• 🎭 Multi-Agent System: 3 AI agents that listen to each other and discuss dev/debug/test topics
• 🌐 Web UI: Beautiful responsive dashboard with live monitoring
• 🧪 Advanced Testing: Complex vocabulary and technical terminology tests
• 📊 Analytics: Performance metrics and accuracy tracking
• ⚡ Fast: 1.4-1.6x speed playback for efficient testing

📦 Installation

Quick Start (npx)

npx 666fr3k --help

Global Installation

npm install -g 666fr3k

Local Installation

npm install 666fr3k

🔧 Prerequisites

Node.js Dependencies

All Node.js dependencies are installed automatically with the package.

Python Dependencies

Install Python dependencies:

# Automatic installation
npx 666fr3k install-deps

# Or manual installation
pip3 install gtts SpeechRecognition pyaudio

System Requirements

• Node.js: v16.0.0 or higher
• Python: 3.7 or higher
• Audio: ffmpeg, ffplay (for audio playback)
• Microphone: Required for STT features

Install System Audio Tools

Ubuntu/Debian:

sudo apt-get install ffmpeg portaudio19-dev python3-pyaudio

macOS:

brew install ffmpeg portaudio

Windows:

choco install ffmpeg

🚀 Quick Start

NEW: Auto-Improvement Loop

The flagship feature - run the complete auto-improvement system:

# Run the FULL auto-improvement loop (recommended)
npx 666fr3k auto-improve

# Run with specific mode
npx 666fr3k auto-improve --mode advanced

# Run simple mode (Python agents only)
npx 666fr3k auto-improve --mode simple --iterations 5

# Start with web dashboard
npx 666fr3k web &
npx 666fr3k auto-improve
# Then open http://localhost:3666

What it does:

1. Monitors all agent conversations in real-time
2. Extracts actionable insights using AI analysis
3. Automatically integrates improvements into:
   • Workflow files (CLAUDE.md)
   • Codebase (relevant source files)
   • Knowledge base (auto-patterns directory)
   • MCP tool configurations
4. Generates unique audio updates (never repeats)
5. Tracks all changes in auto-improvement-state.json

Features:

• Runs continuously (Ctrl+C to stop)
• Saves state every iteration
• Auto-resumes from saved state
• WebSocket-based live monitoring
• Hash-verified unique audio generation

CLI Commands

1. Auto-Listener (Continuous Speech Recognition)

Listen continuously for speech and transcribe in real-time:

# Basic listening (text output only)
npx 666fr3k listen

# With voice responses
npx 666fr3k listen --speak

# Verbose mode
npx 666fr3k listen --speak --verbose

Voice Commands:

• "stop listening" - Exit the listener
• "status" - Show session statistics
• "help" - Show available commands

2. TTS→STT Loop Testing

Run continuous loop tests to verify the pipeline:

# Run 5 cycles at 1.6x speed (default)
npx 666fr3k loop

# Custom cycles and speed
npx 666fr3k loop --cycles 10 --speed 1.8

# Run 20 cycles at normal speed
npx 666fr3k loop -c 20 -s 1.0

3. Verification Tests

Run comprehensive TTS and STT tests:

# Basic tests
npx 666fr3k test

# Advanced tests with complex vocabulary
npx 666fr3k test --advanced

# Custom loop count
npx 666fr3k test --loop 10

4. Multi-Agent Conversation (NEW! 🎭)

Run 3 AI agents that listen to each other and discuss development topics:

# Run 5 rounds of conversation (default)
npx 666fr3k agents

# Custom rounds and speed
npx 666fr3k agents --rounds 10 --speed 1.6

# Short form
npx 666fr3k agents -r 3 -s 1.4

Features:

• 3 agents with different roles (Lead Developer, QA Engineer, DevOps Specialist)
• Agents listen to each other's speech
• Discuss real development/debug/test topics
• Conversational turn-taking
• 1.4-1.6x speed for efficient conversations

5. Web UI

Launch the web interface:

# Start on default port 3666
npx 666fr3k web

# Custom port
npx 666fr3k web --port 8080

Then open: http://localhost:3666

Web UI Features

The web interface provides:

• TTS Panel: Convert text to speech with speed control
• STT Panel: Live microphone transcription
• Loop Testing: Run automated TTS→STT cycles
• Statistics Dashboard: Track usage and accuracy metrics
• Real-time Updates: WebSocket-based live transcription

📚 API Usage

JavaScript/Node.js

const { spawn } = require('child_process');

// Run TTS
const tts = spawn('npx', ['666fr3k', 'loop', '--cycles', '3']);
tts.stdout.on('data', (data) => {
  console.log(data.toString());
});

// Start auto-listener
const listener = spawn('npx', ['666fr3k', 'listen', '--speak']);

Python Integration

import subprocess

# Run loop test
result = subprocess.run(['npx', '666fr3k', 'loop', '--cycles', '5'])

# Start listener
listener = subprocess.Popen(['npx', '666fr3k', 'listen'])

HTTP API (Web Server)

# Start server
npx 666fr3k web --port 3666

TTS Endpoint:

curl -X POST http://localhost:3666/api/tts \
  -H "Content-Type: application/json" \
  -d '{"text": "Hello world", "speed": 1.6}'

STT Endpoint:

curl -X POST http://localhost:3666/api/stt \
  -H "Content-Type: application/json" \
  -d '{"audioData": "<base64_audio>"}'

🧪 Testing Examples

Basic Loop Test

npx 666fr3k loop --cycles 5

Output:

🔄 CYCLE #1
📢 Input: "Hello, this is cycle number one..."
🔊 Playing audio at 1.6x speed... ✓
🎧 Transcribing... ✓
📝 Output: "hello this is cycle number one"
📊 Accuracy: 80.0%
✅ SUCCESS

Advanced Vocabulary Test

npx 666fr3k test --advanced

Tests complex terminology including:

• Technology & AI concepts
• Medical & scientific terms
• Legal & constitutional language
• Quantum physics & mathematics
• Economic & financial theory
• Philosophical concepts
• Cybersecurity terminology

Continuous Listening Session

npx 666fr3k listen --speak

🎤 666FR3K AUTO-LISTENER ACTIVATED
📢 I'm now listening continuously...
💡 Say 'stop listening' to exit

[12:34:56] 🎤 Heard: "Hello, can you hear me?"
💬 You said: Hello, can you hear me?

🎧 Listening...

📊 Performance Metrics

Test Results

• Basic Loops: 80-100% accuracy on simple phrases
• Complex Vocabulary: 16-72% accuracy on technical terms
• Speed: 1.6x playback maintains 70%+ accuracy
• Latency: <2 seconds per TTS→STT cycle

Tested Domains

✅ General conversation (90%+ accuracy) ✅ Technology terminology (80%+ accuracy) ✅ Economic concepts (72% accuracy) ⚠️ Medical terms (37% accuracy) ⚠️ Quantum physics (17% accuracy)

🔧 Configuration

Environment Variables

# Set custom port for web server
export PORT=8080

# Python interpreter path
export PYTHON_BIN=/usr/bin/python3

Package Configuration

Edit package.json to customize:

{
  "scripts": {
    "start": "node index.js",
    "web": "node server.js",
    "test": "node test.js"
  }
}

🐛 Troubleshooting

Microphone Not Working

# Test microphone access
python3 -c "import speech_recognition as sr; print(sr.Microphone.list_microphone_names())"

Audio Playback Issues

# Check ffplay installation
ffplay -version

# Test audio output
ffplay -nodisp test.mp3

Python Dependencies

# Reinstall dependencies
pip3 install --upgrade gtts SpeechRecognition pyaudio

# On macOS, if pyaudio fails:
brew install portaudio
pip3 install --global-option='build_ext' --global-option='-I/usr/local/include' --global-option='-L/usr/local/lib' pyaudio

Google API Rate Limits

If you encounter API timeouts:

• Add delays between requests
• Consider using offline STT alternatives
• Check network connectivity

📖 Command Reference

| Command | Description | Options | |---------|-------------|---------| | auto-improve | 🚀 NEW: Auto-improvement loop | --mode advanced\|simple, --iterations N, --rounds N | | listen | Start continuous listening | --speak, --verbose | | loop | Run TTS→STT loop test | --cycles N, --speed X | | test | Run verification tests | --advanced, --loop N | | agents | 3-agent conversation system | --rounds N, --speed X | | web | Start web UI server with monitoring | --port N | | install-deps | Install Python dependencies | - |

Auto-Improvement Loop Options

# Advanced mode (Full JS monitoring) - DEFAULT
npx 666fr3k auto-improve --mode advanced

# Simple mode (Python agents only)
npx 666fr3k auto-improve --mode simple --iterations 3 --rounds 2

Advanced Mode Features:

• Real-time agent conversation monitoring
• MCP server integration (hey-fr3k, fr3k-think, md-mcp)
• Automatic code and workflow updates
• Hash-verified unique audio generation
• Self-adapting knowledge base
• Runs indefinitely until stopped

Simple Mode Features:

• Python-based multi-agent conversations
• Fixed iteration count
• Lighter resource usage
• Good for testing

🎨 Web UI Screenshots

Main Dashboard:

• TTS Panel with text input and speed control
• STT Panel with live transcription
• Loop testing interface
• Real-time statistics dashboard

🤝 Contributing

Contributions welcome! Please follow these guidelines:

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests
5. Submit a pull request

📝 License

MIT License - see LICENSE file for details

👤 Author

fr3k

• GitHub: @fr3k
• NPM: 666fr3k

🙏 Acknowledgments

• gTTS: Google Text-to-Speech library
• SpeechRecognition: Python speech recognition library
• Google Cloud: Speech API
• Express.js: Web server framework

📈 Roadmap

v2.0 ✅ COMPLETE

• [x] Auto-improvement loop
• [x] Agent conversation monitoring
• [x] AI-powered insight extraction
• [x] Automatic workflow integration
• [x] Dynamic audio generation (never repeats)
• [x] Real-time web dashboard
• [x] MCP server integration

v2.1+ Planned

• [ ] Offline STT support
• [ ] Multiple TTS voice options
• [ ] Docker containerization
• [ ] Multi-language support
• [ ] Custom wake words
• [ ] Voice activity detection improvements
• [ ] Machine learning model fine-tuning
• [ ] Multi-modal insight extraction (code + audio + visual)
• [ ] Distributed agent networks
• [ ] Mobile app

🔥 Quick Examples

Example 1: Quick Test

npx 666fr3k loop --cycles 3

Example 2: Voice Assistant Mode

npx 666fr3k listen --speak

Say "Hello" and hear the system respond!

Example 3: Web Interface

npx 666fr3k web
# Open http://localhost:3666

Example 4: Advanced Testing

npx 666fr3k test --advanced

Tests complex vocabulary across 7 domains!

🔥 Architecture

Auto-Improvement Loop Flow

┌─────────────────────────────────────────────────────────────┐
│                  Agent Conversations                         │
│         (Development, QA, DevOps discussions)               │
└─────────────────────┬───────────────────────────────────────┘
                      │
                      ▼
┌─────────────────────────────────────────────────────────────┐
│              Real-Time Monitoring                            │
│    • File watchers    • Process output capture              │
│    • WebSocket streams • Log aggregation                    │
└─────────────────────┬───────────────────────────────────────┘
                      │
                      ▼
┌─────────────────────────────────────────────────────────────┐
│           AI-Powered Insight Extraction                      │
│  • Pattern recognition  • fr3k-think MCP analysis           │
│  • Confidence scoring   • Priority assignment               │
└─────────────────────┬───────────────────────────────────────┘
                      │
                      ▼
┌─────────────────────────────────────────────────────────────┐
│              Automatic Integration                           │
│  ┌──────────────┬──────────────┬──────────────┐            │
│  │   Workflow   │   Codebase   │   Knowledge  │            │
│  │  CLAUDE.md   │  Auto-fix    │  Patterns    │            │
│  └──────────────┴──────────────┴──────────────┘            │
└─────────────────────┬───────────────────────────────────────┘
                      │
                      ▼
┌─────────────────────────────────────────────────────────────┐
│           Dynamic Audio Generation                           │
│  • Hash verification  • Unique content synthesis            │
│  • Never repeats      • Reflects latest changes             │
└─────────────────────┬───────────────────────────────────────┘
                      │
                      ▼
┌─────────────────────────────────────────────────────────────┐
│              State Persistence                               │
│  auto-improvement-state.json • Progress tracking            │
│  • Resume capability          • Metrics history             │
└─────────────────────────────────────────────────────────────┘

🎓 Use Cases

1. Continuous Development Improvement: Monitor development team conversations and automatically improve workflows
2. AI Agent Research: Study multi-agent communication patterns and emergent behaviors
3. Automated Documentation: Extract insights from conversations and auto-generate documentation
4. Code Quality Enhancement: Automatically identify and apply code improvements
5. Knowledge Base Growth: Build self-expanding knowledge repositories
6. Audio-First Workflows: Generate unique audio updates for accessibility and monitoring

Made with 🔥 by fr3k

Version: 2.0.0 Last Updated: 2025-11-02 License: MIT