as I just wrote in the other post, the /search/series api behaves a bit strange and in parts not according to the documentation while handling languages / translations. Appart from that, which I believe are errors, I'd like to suggest a few improvements:
If a series exists, /search/series should never act like it doesn't, even if there's no matching language available.
If there is no matching language and only a single language available, send that one. The user surely prefers something over nothing and there's no other choice anyway. Most of the time it'll be the correct language anyway. It's rather unlikely that someone adds a Norwegian series, aired only in Norway, and writes the description in Spanish. Also, people interested in this series will almost always understand Norwegian.
If there is no matching language but more than one available, send the original, i.e. for a Swedish language series, send the Swedish dataset (see 4 below), or at least some reasonable default (probably English).
/search/series should return a "language" field, describing which data it chose to send.
/search/series should return a list of available translations.
This saves a few roundtrips, since a client would not have to test for languages using multiple Accept-Language requests, if there's nothing available for the preferred language.
Accept-Language should accept a comma-separated list of languages, sorted by the user from most to least acceptable.
4. (this is for the database itself)
Add one or two language entries to the database, one for the main spoken, original language and possibly a second for all languages used in the series.
Might be a German problem, but the German titles for foreign-language series quite often are outright bullshit. This would greatly help in automatically selecting the original title. Right now, the user may choose to get English titles for US series, but then he gets English titles for French or German series, too, which ain't much better than getting German titles for US or French series.
Last but not least: Thanks for the database!