📜 SCRIBE

The Open Voice Platform

Your personal memory keeper. From the community, for the community.

---

🎯 What is Scribe?

Scribe is the open-source successor to WhisperWriter - evolved into a complete voice automation platform that learns from you.

Think Siri/Alexa/Cortana, but:

• 🔓 Open source - See exactly how it works, modify anything
• 🏠 Local-first - Your data stays on your machine
• 🧠 Actually learns - Gets smarter with every use
• 🔌 Infinitely extensible - Community-driven plugins for everything
• 💰 Free forever - No subscriptions, no corporate lock-in

---

✨ What Makes Scribe Different?

🗣️ Voice-First Everything

"Hey Scribe, switch to Chrome"           → Switches to Chrome
"Scribe, pause Spotify"                  → Pauses music
"Write this down: Meeting notes..."      → Transcribes to active window
"Where did we talk about mortgage rates?" → Finds conversation + context

🧠 It Actually Remembers

• Conversation Memory: Recalls what you discussed and where
• Voice Profile Learning: Adapts to your speech patterns, vocabulary, jargon
• Cross-Device Sync: Learn on desktop, use on laptop - same intelligence
• Context Awareness: Knows which app, URL, or file you were using

📊 Proves Its Value to YOU

Your Scribe Value Report - This Month

Time Saved: 26.4 hours
├─ Typing: 13.9 hours (45,382 words at speaking speed)
├─ Context Switching: 10.4 hours (1,247 voice commands)
└─ Searching: 2.1 hours (23 instant recalls)

💰 Value at your rate: $1,980

Your Improvement:
├─ Accuracy: 87% → 94% (+7%)
├─ Speed: 145 WPM → 168 WPM (+16%)
└─ Errors: -34%

🔌 Day 1 Extensions

• Window Manager: Control windows by voice
• Media Control: Spotify, YouTube, system volume
• Memory Scribe: Remember and recall conversations
• Sync Scribe: Share learnings across devices
• + Community Plugins: Build your own!

⚡ GPU Acceleration (New!)

• 5-10x faster transcription with NVIDIA GPU support
• Automatic GPU detection and fallback to CPU
• Supports all modern NVIDIA GPUs (GTX 10 series+)
• See GPU_QUICKSTART.md for setup

---

🚀 Quick Start

Prerequisites

• Python 3.11+
• Windows 10/11 (Mac/Linux support coming)
• Microphone
• (Optional) API key for AI enhancement

Installation

# Clone the repository
git clone https://github.com/yourusername/scribe.git
cd scribe

# Install in development mode
pip install -e .

# Or use pip when released
pip install scribe-voice

First Run

# Launch Scribe (recommended - if installed)
scribe

# Or run directly from source
python scribe.py

Setup Wizard

On first run, Scribe will:

1. ✅ Check your Python environment
2. ✅ Configure your microphone
3. ✅ Set up your voice profile
4. ✅ (Optional) Add your API key for AI features
5. ✅ Choose your privacy/telemetry level
6. ✅ Enable default plugins

---

🎮 How to Use

Basic Transcription (Classic Mode)

Press: Ctrl+Win (or configured hotkey)
Speak: "Send an email to John about the project update..."
Release: [Scribe types into active window]

Voice Commands (New!)

You: "Hey Scribe"
Scribe: 🎤 Active

You: "Switch to Chrome"
Scribe: ✓ Switched to Chrome

You: "Pause Spotify"
Scribe: ✓ Paused Spotify

You: "Thanks"
Scribe: 💤 Sleeping

Memory & Context (Game Changer!)

[Monday - Browsing mortgage rates in Chrome]
You: "Scribe, remember this page"
Scribe: ✓ Noted

[Friday - Writing code in VS Code]
You: "Scribe, where did we talk about mortgage rates?"
Scribe: "On a Chrome page: 'Top 10 Mortgage Companies'.
        Would you like me to open it?"

You: "Yes"
Scribe: ✓ [Opens exact page]

---

🏗️ Architecture

scribe/
├── src/scribe/
│   ├── core/              # Transcription engine
│   ├── plugins/           # Extension system
│   │   ├── window_manager/    # Control windows
│   │   ├── media_control/     # Control media playback
│   │   ├── memory_scribe/     # Remember conversations
│   │   └── sync_scribe/       # Multi-device sync
│   ├── analytics/         # Value tracking & learning
│   ├── ai/               # AI enhancement
│   └── ui/               # User interface
│
├── data/                 # Your data (gitignored)
│   ├── analytics/        # Voice profile, learnings
│   ├── logs/            # Session logs
│   ├── metrics/         # Performance data
│   └── sessions/        # Conversation history
│
├── profiles/            # Multi-user support
│   └── default/         # Your profile
│
└── plugins/             # Custom/community plugins

---

🔌 Plugin System

Day 1 Plugins

1. Window Manager

Commands:
- "switch to {app}"      → Activate application
- "minimize"             → Minimize current window
- "maximize"             → Maximize current window
- "close window"         → Close current window

2. Media Control

Commands:
- "pause [app]"          → Pause playback
- "play / resume"        → Resume playback
- "next song / skip"     → Next track
- "volume up / down"     → Adjust volume
- "mute"                 → Mute audio

3. Memory Scribe (Opt-in)

Commands:
- "remember this"                    → Save current context
- "where did we talk about {topic}"  → Search conversations
- "open that page"                   → Restore context
- "forget that"                      → Delete memory

4. Sync Scribe

Commands:
- "sync to {device}"     → Sync voice profile
- "export profile"       → Backup profile
- "import profile"       → Load profile from file

Build Your Own Plugin

# plugins/my_plugin/plugin.py
from scribe.plugins.base import BasePlugin

class MyPlugin(BasePlugin):
    name = "my_plugin"
    version = "1.0.0"

    def commands(self):
        return [
            {
                'patterns': ['do something cool'],
                'handler': self.do_something,
                'examples': ['do something cool']
            }
        ]

    def do_something(self):
        return "Something cool done!"

See Plugin Development Guide for details.

---

📊 Analytics & Privacy

Your Data, Your Choice

Scribe tracks analytics to prove its value to you:

• ✅ Time saved
• ✅ Accuracy improvements
• ✅ Command usage
• ✅ Feature effectiveness

All data stays local by default. Optionally share anonymized usage patterns to help improve Scribe.

Privacy Levels

1. None (Default): No data shared, complete privacy
2. Errors Only: Share crash reports to fix bugs
3. Usage Insights: Share feature usage counts (no content)
4. Full Collaboration: Detailed patterns (still no conversation content)

You can view/export/delete all collected data anytime.

---

🌍 Multi-Device Sync

Share your voice learnings across devices:

Option 1: Local Network (Most Private)

# On Desktop
scribe sync start

# On Laptop (same WiFi)
scribe sync discover
scribe sync connect desktop

Option 2: Manual Export (Maximum Privacy)

# Desktop
scribe profile export --output my-profile.scribe

# Laptop
scribe profile import my-profile.scribe

Option 3: Cloud Sync (Convenient, E2E Encrypted)

# Enable cloud sync (you control the keys)
scribe sync enable-cloud

Your API keys are NEVER synced - each device keeps its own.

---

🎯 Roadmap

v2.0 - The Foundation (Current)

• ✅ Voice transcription with AI cleanup
• ✅ Plugin architecture
• ✅ Window Manager plugin
• ✅ Media Control plugin
• ✅ Value analytics
• ⏳ Wake word detection ("Hey Scribe")
• ⏳ Memory Scribe plugin
• ⏳ Multi-device sync

v2.5 - Intelligence (Q1 2025)

• Voice macros (multi-step workflows)
• Context awareness (knows what app you're in)
• Proactive suggestions
• Custom wake words

v3.0 - Platform (Q2 2025)

• Cross-platform (macOS, Linux)
• Plugin marketplace
• Team profiles
• Advanced automation

---

🤝 Contributing

Scribe is open source and community-driven. We welcome:

• 🐛 Bug reports
• 💡 Feature requests
• 🔌 Plugin development
• 📚 Documentation improvements
• 🧪 Testing and feedback

See CONTRIBUTING.md for guidelines.

Why Contribute?

• Build features you want to use
• Learn voice AI, plugin architecture, Python packaging
• Join a community building the open alternative to corporate voice assistants
• Your contributions help everyone

---

📖 Documentation

• User Guide - Complete usage documentation
• Plugin Development - Build your own plugins
• Analytics Guide - Understanding your metrics
• Team Deployment - Set up for your team
• Development Guide - Contributing to Scribe

---

💬 Community

• GitHub Issues: Report bugs, request features
• Discussions: Ask questions, share plugins
• Discord: Join the community (coming soon)

---

📜 License

Scribe is licensed under the Apache License 2.0.

TL;DR: Free to use, modify, distribute. No warranties. Attribution appreciated.

---

🙏 Acknowledgments

Scribe builds on the foundation of:

• WhisperWriter by savbell - The original inspiration
• faster-whisper - Fast, accurate transcription
• OpenAI Whisper - Revolutionary speech recognition
• The open-source community - For making this possible

---

🚀 Why Scribe Exists

Corporate voice assistants:

• Lock you into their ecosystem
• Send your data to their servers
• Prioritize their business goals
• Cost money or show ads

Scribe:

• Works with any tool you use
• Keeps your data local
• Prioritizes YOUR productivity
• Free forever, no strings attached

People deserve better than corporate software. Scribe is the community's answer.

---

📈 Project Status

Current Version: v2.0.0-alpha Status: Active development Stability: Alpha (usable, but expect changes) Looking for: Early adopters, contributors, feedback

---

❓ FAQ

Q: Is this really free? A: Yes. Open source, MIT licensed, no hidden costs.

Q: Does it work offline? A: Yes! Local models work completely offline. AI features require API keys.

Q: Can I use my own API keys? A: Absolutely. Your keys, your control.

Q: What about my privacy? A: All data local by default. You control what (if anything) gets shared.

Q: Can I build commercial products with this? A: Yes! The license allows commercial use.

Q: Why not just use Whisper directly? A: Scribe adds: learning, plugins, analytics, multi-device sync, UI, and community.

---

Built with ❤️ by the community, for the community

⭐ Star on GitHub • 📖 Read the Docs • 🐛 Report Bug • 💡 Request Feature