SCRUM-CLI

AI meeting assistant that transcribes audio and roasts your corporate buzzwords.

What it does

• Transcribes meeting audio with speaker names
• Interactive AI chat about your meetings
• /ridiculous mode roasts meeting culture with buzzword counts
• Vector search across all meetings
• Real-time transcription

Install & Run

Quick Install with uv (Recommended)

# Install uv package manager
curl -LsSf https://astral.sh/uv/install.sh | sh

# Clone and install
git clone https://github.com/Rachit-Gandhi/scrum-cli.git
cd scrum-cli
uv sync
cp .env.example .env
# Add your API keys to .env
uv run scrum-cli transcribe meeting.mp3 --chat

Traditional pip install

pip install scrum-cli
cp .env.example .env
# Add your API keys to .env
scrum-cli transcribe meeting.mp3 --chat

API Keys Needed

Get these and put in .env:

• GEMINI_API_KEY from Google AI Studio
• HUGGING_FACE_TOKEN from HuggingFace

⚠️ HACKATHON JUDGES: Please use your own API keys. The ngrok tunnel below will be updated during judging.

Live Demo Access

Ngrok URL: [WILL BE UPDATED FOR JUDGING]

# This will be replaced with live tunnel URL

Chat Commands

/ridiculous   # Roast mode with buzzword analysis
/summary      # Meeting summary
/actionitems  # Extract action items  
/stats        # Speaker stats
/whosaid      # Search who said what

Example Output

🎭 ROAST ANALYSIS
├── 🎯 Buzzwords: "synergy" (7x), "leverage" (12x)  
├── 🗣️ Mike dominated 67% of airtime
├── 📝 Grade: D+ ("More buzzwords than decisions")
└── 🏆 "Let's leverage synergistic solutions"

⚙️ Requirements

• Python 3.9+
• Required API Keys:
  • Google Gemini API key (required)
  • HuggingFace token (for speaker diarization)

📖 Usage Examples

Basic Transcription

# Transcribe an audio file
scrum-cli transcribe meeting.mp3

# With speaker diarization
scrum-cli transcribe meeting.mp3 --speakers

Interactive Chat Interface

# Start transcription with chat interface
scrum-cli transcribe meeting.mp3 --chat

Available Chat Commands

/help         # Show all commands
/summary      # Professional meeting summary  
/ridiculous   # Hilarious roast analysis 🎭
/actionitems  # Extract action items
/decisions    # Show decisions made
/stats        # Speaker statistics
/whosaid      # Find who said what
/live         # Recent transcription (during streaming)
/export       # Save chat history
/quit         # Exit

Chat Examples

> Who talked the most in the meeting?
🤖 Mike dominated with 67% of speaking time (12.4 minutes)

> /ridiculous
🎭 ROAST MODE ACTIVATED! 
*Generates hilarious analysis with buzzword counts and witty observations*

> What were the main action items?
📋 Found 3 action items:
1. Sarah will update the budget by Friday
2. Mike to schedule follow-up with clients  
3. Team to review proposal draft

🏗️ Architecture

scrum-cli/
├── scrum_cli/              # Main package
│   ├── main.py            # CLI entry point
│   ├── chatbot_ui.py      # Rich TUI chat interface  
│   ├── live_transcription.py  # Real-time audio processing
│   ├── memory_store.py    # Vector database integration
│   ├── proxy_server.py    # FastAPI backend for AI
│   ├── roast_engine.py    # Meeting culture analysis
│   └── ngrok_manager.py   # Tunnel management
└── transcription-tool/     # Audio processing utilities

🎯 Features in Detail

Real-Time Transcription

• Supports multiple audio formats (MP3, WAV, M4A, etc.)
• Advanced speaker diarization using PyAnnote
• Streaming transcription with live updates
• Meeting data persistence with vector search

AI-Powered Analysis

• Professional Mode: Extract action items, decisions, and summaries
• Roast Mode: Hilarious analysis of meeting culture and patterns
• Natural Language Queries: Ask questions about meeting content
• Contextual Memory: Remembers past meetings and speaker preferences

Meeting Analytics

• Speaker time distribution and statistics
• Buzzword detection and counting
• Meeting efficiency scoring
• Personality profile generation
• Export capabilities for further analysis

🔧 Development

Setting up Development Environment

With uv (Recommended)

# Clone repository
git clone https://github.com/Rachit-Gandhi/scrum-cli.git
cd scrum-cli

# Install dependencies and set up virtual environment
uv sync

# Run in development mode
uv run scrum-cli --help

# Run tests
uv run pytest

Traditional setup

# Clone repository
git clone https://github.com/Rachit-Gandhi/scrum-cli.git
cd scrum-cli

# Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install in development mode
pip install -e ".[dev]"

# Run tests
pytest

Project Structure

The project follows a modular architecture:

• scrum_cli/ - Main package with core functionality
• transcription-tool/ - Audio processing and transcription utilities
• Tests and development files are hidden via .gitignore

📦 PyPI Deployment

Note: PyPI upload is currently in progress. For now, please use the local installation method with uv or pip from the source code.

The project is configured for PyPI publication with:

• Proper setup.py and pyproject.toml configuration
• Version management and dependency specifications
• Comprehensive metadata and classifiers
• Development dependencies separation
• GitHub Actions workflow for automated publishing

Sorry for the inconvenience - we're working on getting this published to PyPI soon!

🤝 Contributing

We welcome contributions! Please:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

🆘 Support & Community

• 🐛 Bug Reports: GitHub Issues
• 💬 Feature Requests: GitHub Discussions
• 📧 Contact: Create an issue for support

🎉 Acknowledgments

• Built with Rich for beautiful terminal UI
• Powered by Google Gemini for AI capabilities
• Uses PyAnnote for speaker diarization
• Audio processing with Whisper and Pydub

---

Built with ❤️ for developers who believe meetings should be both productive and entertaining.

Try /ridiculous mode - you won't regret it! 🎭