BCH AI Meeting Notes Generator

An intelligent meeting minutes generator powered by Azure OpenAI and Streamlit. This application transforms meeting transcripts into well-structured, actionable documentation with automated analysis and multiple export formats.

Features

• Multiple Input Formats: Support for TXT, PDF, DOCX, and MD files
• Intelligent Processing: Uses Azure OpenAI for content generation and analysis
• Rich Output Formats: Export as Markdown, DOCX, or PDF
• Meeting Analysis: Automated scoring across multiple dimensions:
  • Meeting Effectiveness
  • Participation & Engagement
  • Action Item Management
  • Risk Management
  • Communication Quality
• Visual Documentation: Automated generation of Mermaid diagrams for:
  • Process flows
  • System architectures
  • Timeline representations
  • Organization structures
  • Project dependencies

Prerequisites

• Python 3.8+
• Pandoc (for document conversion)
• Mermaid-filter (for diagram generation)
• Azure OpenAI API access

Environment Setup

1. Clone the repository:

git clone https://github.com/yourusername/bch-ai-meeting-notes.git
cd bch-ai-meeting-notes

2. Create and activate a virtual environment:

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

3. Install dependencies:

pip install -r requirements.txt

4. Install system dependencies:

# On macOS
brew install pandoc
npm install -g mermaid-filter

# On Ubuntu/Debian
sudo apt-get install pandoc
npm install -g mermaid-filter

# On Windows
choco install pandoc
npm install -g mermaid-filter

5. Create a .env file with your Azure OpenAI credentials:

AZURE_OPENAI_API_KEY=your_api_key
AZURE_OPENAI_API_VERSION=your_api_version
AZURE_OPENAI_API_BASE=your_endpoint
AZURE_OPENAI_DEPLOYMENT_NAME=your_deployment_name

Running the Application

Local Development

streamlit run main.py

Docker Deployment

1. Build the Docker image:

docker build -t bch-ai-meeting-notes .

2. Run the container:

docker run -p 8501:8501 --env-file .env bch-ai-meeting-notes

Access the application at http://localhost:8501

Usage

1. Choose input method:
   • Upload a transcript file (TXT, PDF, DOCX, MD)
   • Paste text directly
2. Click "Generate Minutes"
3. Review the generated minutes and analysis
4. Download in your preferred format (MD, DOCX, PDF)

Project Structure

bch-ai-meeting-notes/
├── main.py              # Main application file
├── schema.json          # Analysis schema definition
├── requirements.txt     # Python dependencies
├── .env                # Environment variables (not in git)
├── Dockerfile          # Docker configuration
├── docker-compose.yml  # Docker Compose configuration
└── images/             # Application images
    ├── bch.png
    ├── user.png
    └── user.svg

License

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