Zendesk Voice Summary Tool

An automated tool for transcribing and summarizing voice recordings from Zendesk tickets using OpenAI's Whisper and GPT models.

Features

• Downloads voice recordings from Zendesk tickets
• Transcribes audio using OpenAI Whisper API
• Generates intelligent summaries with GPT-5
• Posts summaries back to tickets as private comments
• Handles multiple recordings per ticket
• Skip existing transcripts to save time
• Command-line and interactive modes
• Can be packaged as a standalone executable

Installation

Prerequisites

• Python 3.8+
• Zendesk account with API access
• OpenAI API key

Setup

1. Clone the repository:

git clone https://github.com/yourusername/zendesk-voice-summary.git
cd zendesk-voice-summary

2. Create a virtual environment:

python3 -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

pip install -r requirements.txt

4. Set environment variables:

export ZENDESK_DOMAIN="yourcompany.zendesk.com"
export ZENDESK_EMAIL="your-email@company.com"
export ZENDESK_PASSWORD="your-zendesk-password"
export OPENAI_API_KEY="sk-your-openai-api-key"

Usage

Interactive Mode

Run without arguments to enter interactive mode:

python voice_summary.py

Command Line Mode

Process single ticket:

python voice_summary.py 12345

Process multiple tickets:

python voice_summary.py 12345 12346 12347

Process from URL:

python voice_summary.py https://yourcompany.zendesk.com/agent/tickets/12345

Options

• --no-zendesk - Process recordings without posting to Zendesk
• --skip-existing - Skip recordings that already have transcripts

Examples

# Process without posting to Zendesk
python voice_summary.py --no-zendesk 12345

# Skip existing transcripts
python voice_summary.py --skip-existing 12345 12346

Installation as Command Line Tool

Option 1: Wrapper Script

Install the wrapper script to use voice_summary from anywhere:

./scripts/install_voice_summary.sh

Then use:

voice_summary 12345
voice_summary  # Interactive mode

Option 2: Standalone Executable

Build a standalone executable with PyInstaller:

./scripts/build_standalone.sh

This creates a self-contained executable in dist/voice_summary that can be distributed without Python dependencies.

How It Works

The tool operates in three phases:

1. Download & Transcribe: Downloads MP3 files from Zendesk and transcribes them using Whisper
2. Summarize: Creates intelligent summaries with sections for Description, Troubleshooting, and Next Steps
3. Post to Zendesk: Adds the summary as a private comment on the ticket

File Organization

Processed files are saved with this naming convention:

• Audio: ticket{id}_call{call_id}.mp3
• Transcripts: ticket{id}_call{call_id}.txt
• Combined summaries: ticket{id}_combined_summary.txt

When using the wrapper script, files are organized in ~/zendesk-transcripts/.

Configuration

Set these environment variables:

Variable	Description	Required
ZENDESK_DOMAIN	Your Zendesk subdomain (e.g., company.zendesk.com)	Yes
ZENDESK_EMAIL	Your Zendesk account email	Yes
ZENDESK_PASSWORD	Your Zendesk password	Yes
OPENAI_API_KEY	Your OpenAI API key	Yes

Requirements

• Python 3.8+
• requests - For Zendesk API calls
• openai - For Whisper transcription and GPT summarization

Optional:

• pytz - For timezone support (Python < 3.9)
• pyinstaller - For building standalone executables

License

MIT License - see LICENSE file for details

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Acknowledgments

Built with:

• Zendesk API
• OpenAI Whisper
• OpenAI GPT