TransAIscriber

A Python 3.10 tool to transcribe audio files (e.g., iPhone voice memos) and summarize them with action items, attendees, time, and date using OpenAI GPT.

Features

• Transcribe audio using local Whisper (openai/whisper or faster-whisper) or OpenAI Whisper API
• Summarize transcript with GPT (cost-efficient, default: gpt-3.5-turbo)
• Extract action items, attendees, time, and date
• Configurable via CLI or .env config file

Setup (with Anaconda)

1. Clone the repo and create a new Anaconda environment:

   conda create -n transaiscriber python=3.10
   conda activate transaiscriber

2. Install dependencies:

   pip install -r requirements.txt

3. Copy and edit the config file:

   cp config.example.env .env
   # Edit .env to add your OpenAI API key and preferences

Usage

python transaiscriber.py path/to/audio.m4a --method local_whisper --output_dir output

• You can also set options in .env instead of CLI.
• Supported methods: local_whisper, faster_whisper, openai_api

Extending

• Modular code for easy addition of cloud storage or other features.

Requirements

• Python 3.10+
• See requirements.txt