shorts-vid-stacker

Shorts-vid-stacker is a Python-based tool that utilizes openai's transcription, voice generation, and chat completion apis, along with the ffmpeg CLI to create distinct short-form Youtube videos on various content.

The result is fully automatic AI-generated narrated short form content in a form that entices the user to stay and watch:

• The top video is an active video that engages the user
• The bottom video is a passive video in case the user gets bored.
• The narration and text transcription is to further engage the user.

It also has a utility to automatically upload the video to youtube with selenium and chromedriver.

Installation

1. Clone this repo.

2. Install the required dependencies using pip (NOTE: use conda env and python 3.11).

conda env create -f environment.yaml

3. Setup a .envrc file as below (you can use direnv to load these env vars):
   • OPENAI_KEY="user-value"
   • GCS_DEVELOPER_KEY=user-value
   • GCS_CX=user-value

Usage

1. Add keywords in main.py
2. Add input action and passive videos as well as mp3 files in videos/ and audio/ respectively.
3. Run main.py

License

MIT License