TelegramTranscriber

A Telegram bot that transcribes voice messages using FasterWhipser model.

Features

• ✅ Transcribes Telegram voice messages (OGG/OPUS format)
• 🎧 Uses Regolo.ai faster-whisper-large-v3 model (but can change model and AI inference provider)
• 🔐 Securely loads API keys via .env
• 🗑️ Automatically deletes audio files after processing

---

Requirements

• Python 3.8+
• Telegram Bot Token
• Regolo.AI API key (or your inference provider)

---

Setup

1. Clone the repository

git clone https://github.com/regolo-ai/TelegramTranscriber.git
cd TelegramTranscriber

2. Create and activate a virtual environment

python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

3. Install dependencies

pip install -r requirements.txt

4. Set up environment variables

Copy the .env-example file to .env with your keys.

5. Run

python bot.py

---

Docker Setup (Alternative)

1. Build the Docker image

docker build -t telegram-transcriber .

2. Run with Docker

docker run -d --name telegram-transcriber telegram-transcriber

Note: Make sure your .env file is properly configured before building the image.