/GPT-VideoSynthYT

GPT-YouTube-Short-Make is an AI-powered tool leveraging OpenAI's GPT-4 to recreate engaging and motivational videos based on YouTube short links. This project encompasses video extraction, script generation, voiceover synthesis, music recognition, and title creation, culminating in a final composite video.

Primary LanguageJupyter Notebook

GPT-YouTube-Short-Make

GPT-YouTube-Short-Make is an AI-powered tool that utilizes OpenAI's GPT-4 model to recreate videos based on YouTube short links. This project aims to generate engaging and motivational video content by leveraging artificial intelligence.

Overview

The repository contains a set of scripts and notebooks that demonstrate the process of:

  • Extracting audio and video content from a provided YouTube short link.
  • Generating a script using GPT-4 based on a reference transcript.
  • Creating an audio voiceover for the script using OpenAI's text-to-speech (TTS) capabilities.
  • Recognizing music in the video and obtaining a corresponding music track.
  • Adding subtitles to the video with auto-subtitling tools.
  • Generating titles for distinct video clips based on the video's SRT caption.
  • Incorporating Pexels API to obtain relevant video clips for each title.
  • Combining all elements into a final composite video.

Usage

To recreate the provided example, run the main.ipynb notebook. Make sure to set the required environment variables, such as OPENAI_API_KEY and PEXELS_API_KEY.

Sample Output

Check out the sample YouTube channel to view the output generated by this tool.