AITranscribe

A lightweight, fast audio transcription service that automatically converts audio files to text, generates summaries, and saves them as markdown files. Built with Python, Whisper AI, and Gemini AI.

Features

• 🎯 Automatic audio file detection and processing
• 🔄 Automatic file format conversion to MP3 using FFmpeg
• 📝 High-quality transcription using OpenAI's Whisper
• 📚 AI-powered summaries using Google's Gemini
• 📋 Clean markdown output format
• 🔄 Background processing service
• 🚀 FastAPI endpoint for direct uploads

Setup

1. Install dependencies:

pip install fastapi uvicorn whisper google-generativeai python-dotenv python-multipart

2. Install FFmpeg (required for audio conversion)

3. Create a .env file with:

GENERATIVEAI_API_KEY=your_gemini_api_key
SCAN_DIR=path/to/input/folder
TRANSCRIPT_DIR=path/to/output/folder

Usage

Option 1: Automatic File Processing

1. Start the server:

uvicorn main:app --host 0.0.0.0 --port 8000

2. Drop audio files into your SCAN_DIR folder
3. Files will be automatically processed and transcripts will appear in TRANSCRIPT_DIR

Option 2: API Endpoint

Send POST requests to /transcribe/ with audio files:

curl -X POST -F "file=@your_audio.mp3" http://localhost:8000/transcribe/

Output Format

Transcripts are saved as markdown files with:

• Timestamp of generation
• AI-generated summary
• Full transcription text

Requirements

• Python 3.8+
• FFmpeg
• Google Gemini API key