reception-agent

📞 Reception Agent - Multilingual Call Transcription System

A professional Streamlit-based call transcription system that converts audio in any language to English text with AI-powered analysis.

✨ Features

🚀 Quick Start

1. Install Dependencies

pip install -r requirements.txt

2. Configure Environment

Create a .env file:

SUPABASE_URL=your_supabase_url
SUPABASE_KEY=your_supabase_key

3. Run the Application

python -m streamlit run app.py

Open http://localhost:8501 in your browser.

📦 Project Structure

reception-agent/
├── app.py                    # Main Streamlit application
├── auto_processor.py         # Background folder watcher (optional)
├── requirements.txt          # Python dependencies
├── .env                      # Environment variables (create this)
├── .gitignore               # Git ignore file
├── .streamlit/
│   └── config.toml          # Streamlit configuration
├── README.md                # This file
├── DEPLOYMENT.md            # Deployment instructions
└── GITHUB_UPLOAD_STEPS.md   # GitHub upload guide

🌍 Language Support

Set DEFAULT_LANGUAGE in .env to match your primary language:

Language Code
English (US) en-US
Hindi hi-IN
Kannada kn-IN
Tamil ta-IN
Telugu te-IN
Malayalam ml-IN
Spanish es-ES
French fr-FR
German de-DE
Chinese zh-CN

🔧 Optional: Auto-Processor

Run the background processor to automatically transcribe files placed in a folder:

python auto_processor.py

This monitors C:/CallRecordings for new audio files and processes them automatically.

📊 Database Setup

  1. Create a Supabase account at https://supabase.com
  2. Create a new project
  3. Create a table named call_records with these columns:
    • id (int8, primary key)
    • timestamp (timestamptz)
    • filename (text)
    • transcribed_text (text)
    • analysis (text)
    • language (text)

🚀 Deployment

See DEPLOYMENT.md for instructions on deploying to Streamlit Cloud.

📝 How It Works

  1. Audio Input: User records or uploads audio
  2. Speech-to-Text: Uses Google Speech Recognition (free)
  3. Translation: Automatically translates to English using deep-translator
  4. Analysis: Analyzes intent, sentiment, and action items
  5. Storage: Saves to Supabase database
  6. Display: Shows transcription and analysis results

🛠️ Technologies Used