The short answer

German is the language where people who know the grammar best often speak the least, because the case system punishes a wrong guess and the verb sits at the end so you have to commit early. What helps is talking, badly, until the commitment stops being frightening. The tutor hears your audio rather than a transcript, so it can hear a ü flattened to oo or a ch avoided entirely. If you have a trip booked, Travel Mode rehearses you, sets one real task a day in the city you are in, and debriefs you afterwards.

German has an unusual failure mode. The people who struggle to speak it are often the ones who studied it most carefully, because they have learned enough to know that a wrong article is wrong, and not yet enough to produce the right one at speed.

The verb at the end is the real problem

Most languages let you start a sentence and work out where it is going. German often does not. Subordinate clauses put the verb at the end, the perfect tense puts the participle at the end, and modal constructions stack up behind it. So you have to know roughly where you are heading before you open your mouth.

That is a planning problem, not a knowledge problem, and it only gets better by doing it. Sessions here are unscripted conversation, so you commit, you get it wrong, you recover, and the recovery is the thing that gets faster.

Cases, without the paralysis

Four cases and three genders means a lot of chances to be wrong, and the usual result is that people slow down to compute and lose the thread.

The tutor’s job is not to interrupt every dative you fumble. It is to keep the conversation moving and feed the correct form back inside its own reply, the way a person would, so you hear it in place rather than being marked on it. The patterns you keep missing are tracked and quietly shape later sessions.

Why it helps that the tutor hears you

Your audio goes to the model directly rather than being transcribed into text first, so pronunciation, hesitation and pace are part of the reply.

This matters more in German than people expect. A ü flattened to oo transcribes as ü. A ch replaced with a k transcribes as ch. Final devoicing skipped transcribes as the correct word. Everything you are getting wrong arrives at a text-based tutor already corrected, which is why so many learners have fluent grammar and an accent nobody ever mentioned.

Going to Germany, Austria or Switzerland

Travel Mode changes the app for the length of a trip. Rehearsal before you go, then one real-world mission a day while you are there, set in the city you are in and pitched at or below your level. You rehearse it with your tutor, close the app, go and do it, then come back and tell them what happened.

Worth knowing before you land: everybody will switch to English the moment you hesitate, and it is meant kindly. Having rehearsed the specific exchange that morning is most of what lets you hold the line for the ten seconds it takes for them to decide you are worth speaking German to.

The syllabus behind it runs 321 items from A1.1 to C1, of which 105 are capabilities, concrete things you can do rather than grammar points.

The five German tutors

Every language has five, each from a different place and each sounding like it. You can switch whenever you like, and the app remembers where you are regardless of who you talk to.

LukasBerlin
Standard High German with a relaxed, dry urban delivery.
HannahMunich
Standard German with the warm, slightly sing-song colour of Bavaria.
MaximilianVienna
Standard German spoken with the soft, melodic, courteous lilt of Austria.
GretaHamburg
Crisp northern Hochdeutsch with the clear, slightly cool maritime cadence of the Hanseatic north.
SophieCologne
Friendly Rhineland Hochdeutsch with a soft, sing-song lilt and an open, cheerful warmth.

What English speakers usually find hard

These are drawn from the pronunciation patterns the tutor listens for in German. Because it hears your voice rather than a transcript of it, they are things it can actually notice and work on with you.

ü and ö, which English simply does not have: Say ee and round your lips without moving your tongue, and you have ü. Say eh and round your lips, and you have ö. English speakers default to oo and uh, which turns Tür into Tur and schön into shern. Both are learnable in minutes and most people are never told the trick.
One ch spelling, two different sounds: After front vowels and after n, l or r it is the soft ich sound, like a hissing cat. After a, o, u and au it is the hard ach sound, a guttural scrape. Ich and ach are the training pair, and English speakers usually substitute a k for both.
Voiced consonants go voiceless at the end: Hund is pronounced Hunt, Tag is Tahk, Berg is Berk. The spelling keeps the voiced letter and the sound does not. The plural restores it, so Hunde has a real d, which is why the rule feels inconsistent until you see it.
st and sp become sht and shp at the start of a word: Stadt is shtatt, sprechen is shprechen, Spiel is shpeel. Inside a word they behave normally, so Fenster keeps its st. Also, w is an English v and v is usually an English f, so Wein is vine and viel is feel.
Stress on the first syllable, and prefixes that move it: The default is the first syllable of the stem. Separable prefixes take the stress, which is how you hear that a verb is separable: AUFstehen, ANrufen. Inseparable prefixes do not: beSUCHEN, verSTEHEN. Getting it wrong is immediately audible.