Claude Recording Processor

Automated system for processing Claude recordings from Google Drive. This tool:

1. Monitors a Google Drive folder for new Claude recordings
2. Filters recordings by room ID
3. Applies studio-quality audio processing
4. Generates AI-powered show notes from transcripts
5. Creates segmented episodes from identified conversations
6. Organizes all output files automatically

Features

• Google Drive Integration: Automatically monitors a specified folder for new recordings
• Audio Enhancement: Applies professional studio sound processing using FFmpeg
• AI Show Notes: Uses Claude AI to analyze transcripts and generate descriptive show notes
• Smart Segmentation: Identifies Q&A conversations and creates separate episode files
• Minimum Duration Filtering: Only creates episodes for conversations longer than 10 minutes (configurable)
• Automated Organization: Outputs processed files and segmented episodes to separate folders

Prerequisites

Required Software

1. Node.js (v14 or higher)

   # Check your version
   node --version

2. FFmpeg (for audio processing)

   # macOS
   brew install ffmpeg
   
   # Ubuntu/Debian
   sudo apt-get install ffmpeg
   
   # Windows (using Chocolatey)
   choco install ffmpeg

Required API Access

1. Google Drive API Credentials

   • Go to Google Cloud Console
   • Create a new project or select existing one
   • Enable Google Drive API
   • Create Service Account credentials
   • Download the JSON credentials file
   • Save it as config/google-credentials.json

2. Anthropic API Key

   • Sign up at Anthropic Console
   • Generate an API key
   • You'll add this to your .env file

Installation

1. Clone or navigate to the project directory

   cd claude-recording-processor

2. Install dependencies

   npm install

3. Set up configuration

   # Copy the example environment file
   cp .env.example .env

4. Edit the .env file with your settings

   # Open in your editor
   nano .env

   Required settings:

   • GOOGLE_DRIVE_FOLDER_ID: The ID of your Google Drive folder to monitor
   • GOOGLE_CREDENTIALS_PATH: Path to your Google credentials JSON file
   • ANTHROPIC_API_KEY: Your Claude API key
   • TARGET_ROOM_ID: The recording room ID to filter

5. Place your Google credentials file

   # Save your google-credentials.json to the config folder
   cp /path/to/your/credentials.json config/google-credentials.json

Configuration

Environment Variables

Edit .env to customize:

# Google Drive API credentials
GOOGLE_DRIVE_FOLDER_ID=your_folder_id_here
GOOGLE_CREDENTIALS_PATH=./config/google-credentials.json

# Claude API
ANTHROPIC_API_KEY=your_anthropic_api_key_here

# Target Recording Room ID
TARGET_ROOM_ID=your_room_id_here

# Processing Options
MIN_SEGMENT_DURATION_MINUTES=10
POLL_INTERVAL_MINUTES=5

# Output Directories
PROCESSED_OUTPUT_DIR=./output/processed
SEGMENTED_OUTPUT_DIR=./output/segmented
TEMP_DIR=./temp

Finding Your Google Drive Folder ID

The folder ID is in the URL when you open the folder in Google Drive:

https://drive.google.com/drive/folders/[FOLDER_ID_HERE]

Usage

Continuous Monitoring Mode

Run the processor continuously to monitor for new recordings:

npm start

The system will:

• Check for new recordings every 5 minutes (configurable)
• Process each new recording automatically
• Continue running until you stop it (Ctrl+C)

One-Time Processing

Process all new recordings once and exit:

npm run start:once

Useful for:

• Testing your setup
• Running as a scheduled job (cron, etc.)
• Processing a backlog of recordings

Output Structure

Processed Folder (output/processed/)

Contains the main processed recording with studio sound:

• [recording_name]_studio.mp3 - Enhanced audio file
• [recording_name]_show_notes.md - Complete show notes for all conversations

Segmented Folder (output/segmented/)

Contains individual episode segments:

• episode_1_[conversation_title].mp3 - Audio segment
• episode_1_[conversation_title]_show_notes.md - Episode-specific show notes
• _episodes_summary.md - Overview of all created episodes

How It Works

Processing Pipeline

1. Monitor: Checks Google Drive folder for new files
2. Filter: Only processes files from the specified room ID
3. Download: Downloads audio and transcript files
4. Enhance: Applies studio sound processing to audio
5. Analyze: Uses Claude AI to identify conversations and generate show notes
6. Segment: Creates separate episode files for Q&A conversations over 10 minutes
7. Organize: Saves all output files to appropriate folders
8. Cleanup: Removes temporary files

Audio Processing

The studio sound processing includes:

• High-pass filter (removes low-frequency rumble)
• Noise reduction
• Dynamic range compression (consistent volume)
• EQ adjustments (voice clarity)
• Loudness normalization (-16 LUFS podcast standard)
• Limiter (prevents clipping)

Show Notes Generation

Claude AI analyzes the transcript to:

• Identify distinct conversations (typically 2-3)
• Determine if conversations are Q&A format
• Extract key topics discussed
• Estimate conversation duration
• Generate descriptive paragraphs for each conversation

Episode Segmentation

Creates individual episodes when:

• Conversation is in Q&A format
• Duration is ≥10 minutes (configurable)
• Has clear topic boundaries

Troubleshooting

FFmpeg not found

# Verify FFmpeg is installed
ffmpeg -version

# If not installed, install it (see Prerequisites)

Google Drive API errors

• Verify your credentials file is in the correct location
• Check that the Google Drive API is enabled in your project
• Ensure the service account has access to the target folder

Anthropic API errors

• Verify your API key is correct in .env
• Check your API quota/limits
• Ensure you have internet connectivity

No recordings found

• Verify the GOOGLE_DRIVE_FOLDER_ID is correct
• Check that files contain the target room ID
• Confirm recordings haven't already been processed (check temp/processed-files.json)

Episodes not being created

• Check that conversations are longer than MIN_SEGMENT_DURATION_MINUTES
• Verify conversations are identified as Q&A format in show notes
• Review the console output for filtering messages

Advanced Usage

Changing Minimum Episode Duration

Edit .env:

MIN_SEGMENT_DURATION_MINUTES=15  # Now requires 15+ minutes

Adjusting Polling Interval

Edit .env:

POLL_INTERVAL_MINUTES=10  # Check every 10 minutes instead of 5

Customizing Audio Processing

Edit src/audioProcessor.js to modify the FFmpeg filters in the applyStudioSound method.

Reprocessing Files

If you want to reprocess a file that was already handled:

1. Edit temp/processed-files.json
2. Remove the file ID you want to reprocess
3. Run the processor again

Project Structure

claude-recording-processor/
├── config/
│   ├── config.js              # Configuration loader
│   └── google-credentials.json # Google API credentials
├── src/
│   ├── index.js               # Main orchestration
│   ├── driveMonitor.js        # Google Drive integration
│   ├── audioProcessor.js      # Audio enhancement
│   ├── transcriptProcessor.js # AI show notes generation
│   └── audioSegmenter.js      # Episode segmentation
├── output/
│   ├── processed/             # Studio audio & show notes
│   └── segmented/             # Individual episodes
├── temp/                      # Temporary files during processing
├── .env                       # Your configuration
├── .env.example              # Configuration template
├── package.json              # Node.js dependencies
└── README.md                 # This file

License

ISC

Support

For issues or questions:

1. Check the troubleshooting section above
2. Review console output for error messages
3. Verify all prerequisites are installed correctly
4. Ensure API credentials are valid and have proper permissions