In Czech (and many other languages like Slovak) you have accented characters, but people tend to search without accented characters as it's faster and is unique enough.
So instead of searching belgická 18, one would type belgicka 18, but this is where it gets unreliable.
Search for: !m belgicka 18
The first result for me is in Prague, but yields a result in German and clicking it will return NO POI found.
The third suggestion is in Náchod, and also returns not found.
It happens only sometimes. E.g. Celakovskeho 1120 is broken, Dlouha 10 is not.
Ideally, the results would work the same when you enter both the full name and without accents.
Also, detect the language correctly. After all, Belgicka is closer to Belgická than to Belgischer.
