CropBarsTelegramBot

A Telegram bot that automatically detects and removes black bars (letterboxing/pillarboxing) from videos using FFmpeg. It features an asynchronous worker queue to handle processing-heavy tasks efficiently.

Features

• Automatic black bar detection using FFmpeg's cropdetect filter.
• Support for custom padding: Top, Bottom, Both, or None.
• Queue-based processing with a configurable number of workers.
• Per-user concurrency limits to prevent resource exhaustion.
• Built-in health check server for compatibility with hosting platforms like Render.
• Handles videos sent directly or via the /crop reply command.

Usage

1. Send a video or document directly to the bot.
2. Alternatively, reply to an existing video with the /crop command.
3. The bot will present an inline keyboard asking for the padding position:
   • No Text: Strict crop to detected content.
   • Top: Adds 30px padding to the top (useful for UI).
   • Bottom: Adds 30px padding to the bottom (useful for subtitles).
   • Both: Adds 30px padding to both top and bottom.

Commands

Command	Description
/start	Display welcome message
/crop	Reply to a video with this command to initiate cropping

Configuration

The bot is configured via environment variables.

Variable	Description	Default
BOT_TOKEN	Your Telegram bot token from @BotFather	Required

Self-Hosting

Requirements

• Python 3.x
• FFmpeg installed and accessible in the system PATH.
• A public-facing port (if using health checks for monitoring).

Installation

pip install -r requirements.txt

Running the Bot

Bash

export BOT_TOKEN="your_bot_token_here"
python main.py

The bot runs in polling mode but also starts a web server on 0.0.0.0:7860 for health checks.

---

Deployment to Render

1. Create a new Web Service on Render.
2. Connect your repository.
3. Set Runtime to Python.
4. Set the Start Command to:

  python main.py

5. Add BOT_TOKEN to the Environment Variables.
6. Render will automatically detect the server on port 7860 to verify the service is live.

---

Technical Details

Framework

• aiogram 3.x for asynchronous Telegram interaction.

Processing Engine

• FFmpeg

Workflow

1. The bot downloads the video to a temporary directory.
2. It runs a cropdetect pass to determine the bounding box of the content.
3. It applies user-selected padding adjustments to the crop parameters.
4. It re-encodes the video using libx264 (fast preset) while copying the audio stream to preserve quality.
5. The processed file is uploaded back to the user, and temporary files are cleaned up.

---

Dependencies

• aiogram — Telegram Bot framework
• aiohttp — Asynchronous HTTP server and client
• FFmpeg — Video processing and analysis tool