Chop - Automatic Viral Video Clipper

chop.py is a Python script that automates the process of finding and clipping viral moments from YouTube videos. It uses AI to identify engaging segments and then extracts them into short, shareable clips.

How it Works

The script performs the following steps:

1. User Input: It prompts the user for a YouTube URL and the desired number of clips to generate.

2. Download Video: It takes the YouTube URL as input and downloads the corresponding video using youtube-dl.exe.

3. Download Subtitles: It then downloads the English subtitles for the video in SRT format using yt-dlp.

4. Find Viral Moments with Gemini AI: The core of the application lies in its use of the Gemini AI. It sends the video's subtitles to the Gemini API and asks it to find a specified number of "viral moments". The script provides the following criteria to the AI for what constitutes an engaging moment:

   • Emotional reactions (laughter, surprise, anger, excitement)
   • Surprising facts or insights
   • Strong or controversial opinions
   • Clear questions with punchy answers
   • Quotable lines that stand on their own

   The AI is instructed to find clips that are 10-30 seconds long.

5. Save Clip Information: The start times, end times, and brief reasons for why each moment is engaging (as determined by the AI) are saved into a file named viral_clips.txt.

6. Extract Clips: Finally, the script uses ffmpeg to cut the identified segments from the original video file. The resulting clips are saved as clip_1.mp4, clip_2.mp4, and so on.

Dependencies

The script relies on the following external tools and libraries:

• youtube-dl.exe
• yt-dlp
• ffmpeg
• google-generativeai (Python library)
• python-dotenv (Python library)

Usage

To run the script, you need to have Python installed, along with the necessary dependencies. You will also need to set up a GEMINI_API_KEY environment variable with your API key for the Gemini AI service.

Then, you can run the script from your terminal:

python chop.py

The script will prompt you to enter a YouTube URL and how many clips you want. After you provide the information, it will execute the steps outlined above. If successful, it will create a directory named after the video's title in the 'output' folder. Inside this directory, you will find a separate folder for each clip (e.g., clip_1, clip_2), each containing the video clip and its corresponding SRT subtitle file.