#Hi! I'm trying autoassignment, but I

1 messages · Page 1 of 1 (latest)

queen finch
#

Hi there,
we're basically using LLMs to analyse the script and figure out the characters from the context.
The idea is that when you upload the epub, after a longer processing by LLMs you will get project with text split between characters. Then you can replace voices to desired ones.
It should speed up workflows by not having to select manually all the text said by characters.
But of course it's not 100% correct so you might still need to manually fix some assignments.

#

If you write,
CHARACTERNAME: Charfacter line
it should have fairly high success rate, but for now you would then have to go line by line and remove CHARACTERNAME:
everywhere.