VEWorkflowAutomation

An automation project where we aim to be able to create a full generator of A.I. generated content.

The final product must be able to produce a non-copywritable video consisting of the same content as an original movie with less than 5 second cuts of and ai generated voice over synopsis.

Break Down af actual flow:

1. generate audio synopsis mp3
2. cut movie to length of audio synopsis
   1. remove audio from move
   2. make sure that the video cuts are less than 5 seconds(to avoid being copyright striked)
   3. make sure that length of output video is the same as the length of audio synopsis
3. stitch generated audio synopsys and mp3 into one mp4 file ready for production

Solved problems:

• step 1: generated from a LLMs and ALMs
• step 2.1: solved by ffmpeg

  ffmpeg -i <moviefilename>.mp4 -an -c:v copy <outputfilename>.mp4

• step 3: solved by ffmpeg(not implemented)

problem1

cutting video takes too much time for the app to be viable

proposal1(currently being implemented as Skim)

should have a source of the 5 second clips as mp4 files in the correct order. This way we can programmatically stitch the in a sequence using filename numbers

proposal2(IMPLEMENTED)

manually watching the movie and and listening to the audio file bit by bit.