Yes, they should be providing users with a "starting point" and then syncing from there. If they can't do that, then they should at least start from the monthly JSON dump, use that to populate a local cache, and then sync from there.
Then all queries can hit the local DB instead of the API.
We have to be really careful especially for the Japanese naming conventions. I prefer Romaji for the Katakana but shared words with common spellings, that gets tricky. If the user forgets to fill out the alternate title card with something that DOES say Zelda, then searches for Zelda won't yield the...
The scrapers shouldn't be targeting an app that uses the Api directly like that. They should collating all the data in some way, and using the update API to keep the db in sync. This could mean pushing an entire db to users (like the apps that use the API do) or them writing their own API.