We make videos for youtube using invideo. It convert our text into a script and make a cut using stock images and free video corresponding to descriptions, with voice over.
But the stock are not great. Nor the voice (Eleven Lab is much better) So here is the idea : making an automation on N8N with our own source, and the ability to choose at each step between possibilites.
We are looking for a freelancer to build that
1/ First, an automation to compose a stock
a) Archivebox grabs opensource video from archive.org (movies) with determined sources
b) GPT analyse and describe the scenes to make an information database. this is the most important step
2/ the second automation make the video
a) User input a text and get a script proposition (GPT). User is asked at each step to give confirmation or choose betweens options
b) For each scene (or sequence), GPT make a proposition of content description, and offer a choice according to the video stock. User give confirmation, or input another description.
If not satisfactinf, user can point out sources for grabbing (youtube) or upload a video to cut…
c) An online video plateform cut and assemble the video according to script and sources.
d) the voiceover is generated with through Eleven Labs API
e) Music is generated or imported and added. User give confirmation or choose another one
etc…it’s mor or less the process