>>106315462
>Is there a scraper extension like with szuru for a local danbooru instance
You gotta write that stuff on your own. I know that danbooru has a scraping account by default which is also what gelbooru uses but there are no docs about it.
>and is it easier or harder to do api calls?
Depends. There is pybooru which if it works makes things very easy but it doesn't for me so i make my own API calls. There are some confusing things but its doable. Mainly you first upload some media, get its media ID in the reply, and then make it a post by supplying it with this media ID and the tags and other stuff you want. In general its not as intuitive as szuru.
Also YOU NEED to castrate its auto-tagger. It will hog a lot of resources when you upload media. Remove its address like in this image.
Also modify the docker-compose file to store tumbnails on an SSD and the actual files on an HDD if you want. You should already use ZFS so the HDD is not actually used when its avoidable.

Again, no docs and not very intuitive. Learned these things by trial and error.