Mr.Transcribe - Audio to Text Transcription Web App

A modern web application that transcribes audio files to text with AI-powered refinement capabilities and messaging platform integration.

Features

• 🎤 Multi-format Audio Support: Upload audio files in any format (MP3, WAV, M4A, FLAC, etc.)
• 🌍 High-Quality Transcription: Uses OpenAI Whisper API OR free local Whisper model (automatic fallback)
• 💰 100% FREE Option: Local Whisper transcription runs entirely on your machine - no API costs!
• 💬 AI Chat Interface: Chat with AI to refine, edit, and improve transcriptions
• 📄 Multiple Export Formats: Export transcriptions as TXT, DOCX, PDF, JSON, SRT, VTT
• 📱 Messaging Integration: Connect via Telegram, WhatsApp, or Messenger as "Mr.Transcribe"
• 🔄 Iterative Refinement: Keep refining transcriptions until you're satisfied

Tech Stack

• Frontend: React, TailwindCSS, Vite
• Backend: Node.js, Express
• Transcription: OpenAI Whisper API (with automatic fallback to FREE local Whisper)
• AI Chat: OpenAI GPT-4 or Groq (Llama 3.1 70B)
• Export: docx, pdf-lib, json

Setup

Prerequisites

• Node.js 18+ and npm
• OpenAI API key (required for transcription)
• Groq API key (optional, for chat/refinement - will be preferred if provided)

Installation

# Install dependencies
npm install

# Set up environment variables
cp .env.example .env
# Edit .env with your OpenAI API key

# Start development server
npm run dev

Environment Variables

Create a .env file in the root directory:

# Required for transcription
OPENAI_API_KEY=your_openai_api_key_here

# Optional: For chat/refinement (will be preferred over OpenAI if provided)
GROQ_API_KEY=your_groq_api_key_here

# Server Configuration
PORT=3000
NODE_ENV=development

Note:

• OPENAI_API_KEY is optional - if not provided or quota exceeded, the app automatically uses FREE local Whisper transcription
• GROQ_API_KEY is optional but recommended for faster/cheaper chat responses
• If both API keys are provided, Groq will be used for chat/refinement, OpenAI for transcription
• FREE Mode: If OpenAI quota is exceeded, the app automatically falls back to local Whisper (first use downloads ~500MB model, then works offline forever!)

Deployment

Deploy to mrtranscribe.6x7.gr using Vercel, Netlify, or your preferred hosting.

License

MIT License - see LICENSE file for details.