Published 2025-08-07 by Josue Valentin
This guide describes implementing automated call transcription in FreePBX systems using OpenAI's Whisper engine, operating entirely offline without external APIs or data transmission.
Overview
The solution includes two main components:
- setup_transcriber.sh: Installs Whisper in a clean, self-contained environment without affecting FreePBX or system packages
- transcribe_watcher.sh: Monitors call recordings and automatically transcribes new files
Benefits
- Runs entirely on your own infrastructure
- No external API calls or data transmission
- Keeps sensitive call audio private
- No per-minute transcription costs
Requirements
- FreePBX system with call recording enabled
- Python 3.8 or later
- Sufficient CPU/RAM for transcription processing
- Storage for transcription output files
Security Considerations
Running Whisper locally ensures sensitive call audio never leaves your infrastructure. This addresses privacy concerns for businesses requiring confidential call handling in healthcare, legal, and financial industries.
For detailed implementation instructions, contact our team for assistance with your FreePBX deployment.