🎤 Voice Note Transcription & Notion Integration

A powerful command-line tool for transcribing voice notes and automatically generating structured meeting minutes in Notion. Perfect for engineers, project managers, and professionals who want to streamline their meeting documentation workflow.

---

🌊 About Flocode

This tool is part of Flocode's open-source initiative to empower civil and structural engineers with practical AI-powered tools. As a free community resource, you're welcome to take this tool, modify it, and make it your own! 🚀

---

✨ Features

Feature	Description
🎤 High-Quality Transcription	Utilizes Groq (whisper-large-v3-turbo) and OpenAI Whisper for accurate speech-to-text conversion
🤖 AI-Powered Summarization	Leverages Google Gemini (with OpenAI fallback) to generate structured meeting minutes
📝 Notion Integration	Automatically creates formatted entries in your Notion database with proper markdown rendering
🔔 Real-Time Notifications	Desktop notifications keep you informed of processing status
💾 Local Backups	Maintains local markdown backups of every transcription for your records
🔒 Secure Configuration	All API keys managed securely through environment variables
🚀 Drag & Drop Processing	Simply drag audio files onto the batch script for instant processing
📊 Queue Management	Add multiple files and URLs to a processing queue for batch operations

🚀 Getting Started

📋 Prerequisites

Before you begin, ensure you have the following installed:

• Python 3.12+
• uv: Fast Python package installer

  pip install uv

• FFmpeg: Required for audio processing

📦 Installation

1. Clone the repository:

git clone https://github.com/your-username/whisper_2.0.git
cd whisper_2.0

2. Install dependencies:

uv sync

3. Install FFmpeg:

🪟 Windows (with Chocolatey)

choco install ffmpeg

🍎 macOS (with Homebrew)

brew install ffmpeg

🐧 Linux (with apt)

sudo apt update && sudo apt install ffmpeg

⚙️ Configuration

1. Create your .env file:

cp .env.example .env

2. Configure your API keys:

Service	Environment Variable	Purpose	Required
	OPENAI_API_KEY	Transcription & Summarization fallback	✅
	GROQ_API_KEY	Fast transcription (recommended)	🔄
	GEMINI_API_KEY	Enhanced summarization	🔄
	NOTION_API_KEY	Database integration	✅
	NOTION_DATABASE_ID	Target database	✅
🏢	COMPANY_NAME	Personalized minutes	⚪
🏢	COMPANY_SHORTHAND	Company abbreviation	⚪

Legend: ✅ Required | 🔄 Optional (recommended) | ⚪ Optional (nice-to-have)

🧪 Test Your Setup

Verify everything is configured correctly:

uv run python tests/test_voice_system.py

💻 Usage

🎯 Quick Start (Recommended)

Drag & Drop Processing:

1. Create a desktop shortcut to quick_process.bat
2. Drag your audio file onto the shortcut
3. Get notified of successfull completion ✨

🔄 Interactive Mode

Perfect for managing multiple files:

uv run python scripts/process_voice_notes.py --interactive

Available commands:

• add <file_or_url> - Add to processing queue
• queue - Show current queue
• process - Process next item
• p - Process all items
• clear - Clear queue
• Direct file paths work too!

⚡ Command-Line

For direct processing:

uv run python scripts/process_voice_notes.py /path/to/your/audio.mp3

---

🎯 My Preferred Workflow

🎙️ Recording Setup with VoiceMeeter Banana

Step 1: Audio Capture

• Use VoiceMeeter Banana to record both desktop audio and microphone input into a single audio file for transcription.

Step 2: Instant Processing

1. Create a desktop shortcut to quick_process.bat
2. After your meeting ends, drag the audio file directly onto the shortcut
3. The script handles everything automatically:
   • ✅ Transcribes the entire conversation
   • ✅ Generates structured meeting minutes
   • ✅ Saves to Notion with proper formatting
   • ✅ Creates local markdown backup
   • ✅ Sends you a completion notification

🎛️ Alternative: GUI File Selection

For a more traditional approach, use select_and_process.bat to browse and select files through a Windows dialog.

⚖️ Legal Notice: Always obtain proper consent before recording conversations. Comply with local laws and regulations.

🔧 How It Works

graph TD
    A[📁 Audio File] --> B[🔄 Queue Management]
    B --> C[✂️ Audio Chunking]
    C --> D[🎤 Transcription<br/>Groq/OpenAI Whisper]
    D --> E[🤖 AI Summarization<br/>Gemini/GPT-4]
    E --> F[📝 Notion Integration<br/>Formatted Markdown]
    E --> G[💾 Local Backup<br/>Markdown File]
    F --> H[🔔 Success Notification]
    G --> H

Loading

Process Flow:

1. 📋 Queue Management - Files and URLs organized in processing queue
2. 🎵 Audio Processing - Files chunked for optimal API handling
3. 📝 Transcription - High-quality speech-to-text conversion
4. 🧠 Summarization - AI-powered meeting minutes generation
5. 💾 Dual Storage - Notion database + local markdown backup

---

🤝 Contributing

We welcome contributions from the engineering community! This is an open-source Flocode initiative.

• 🐛 Found a bug? Open an issue
• 💡 Have an idea? Start a discussion
• 🔧 Want to contribute? Submit a pull request

📜 License

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

Free for commercial and personal use. 🎉

---

🌊 Built with ❤️ for the Flocode Community

Empowering engineers with practical AI tools, one voice note at a time.

James 🌊

---