A short explanation of the approach behind the app — for teachers and parents deciding whether to use it, and anyone curious how it's built to actually work.
Full mastery of all 50 U.S. state capitals — not just memorized text pairs, but capitals tied to a real sense of where each state actually is, reinforced through sight, sound, and touch, and built to survive a distracted, low-attention environment like the back seat of a car.
Multiple-choice questions get criticized for letting a student guess by process of elimination instead of actually knowing the answer. This app takes the opposite view: elimination, done right, is a real learning method — reasoning through "that's Oregon's capital, that's California's capital, none of those are this state" teaches three capitals in the process of finding a fourth. The three levels below don't hide that mechanic; they gradually take away the ability to lean on it, until the final level requires knowing the answer outright.
Wrong answers are other states' actual capitals, drawn from anywhere in the country. A student can reason their way to the right answer by recognizing what the other three options are — and learns those three capitals as a side effect.
Wrong answers mix real state capitals with well-known cities that aren't capitals, usually from the same region. Elimination still works, but the student has to keep straight which famous cities actually are capitals.
Wrong answers are well-known cities from the same state as the question. All four options are genuinely in that state, so "which state is this city in" — the trick that carried the first two levels — provides no information at all. The student has to specifically know which city is the seat of government.
This level exists for exactly the states where it matters most: a state whose most famous city (Los Angeles, New York City, Anchorage, Burlington) is not its capital, so a student who has only ever reasoned regionally gets caught guessing the famous city instead of the actual capital.
The Hard-mode city list enforces one rule: a city's name is never used as a wrong answer anywhere in the app if that exact name is also some state's actual capital — even a different state's capital, and even when the two are genuinely different cities that just happen to share a name.
The clearest example: Charleston, South Carolina is a large, famous city, and it is not South Carolina's capital — Columbia is. But Charleston is West Virginia's capital. If "Charleston" showed up as a wrong answer while quizzing South Carolina, the app would be implicitly teaching "Charleston isn't a capital" — true for South Carolina, false in general, and there's no way for a student to know from the option alone which Charleston is meant. So Charleston is excluded from South Carolina's list entirely, even though it's the state's most recognizable city. The same rule quietly affects a handful of other states (Georgia loses Augusta and Columbus, Missouri loses Springfield, Wyoming loses Jackson) — each one a real US city whose name collides with a different state's capital.
Not every student learns a mnemonic the same way. Some genuinely benefit from being handed a ready-made memory hook — it gives them something concrete to picture immediately, with no extra effort before it starts paying off. Others retain almost nothing from a hook someone else invented; for them, the act of inventing their own silly association is what actually makes it stick, because the memory is tied to their own reasoning rather than borrowed words.
The app doesn't force a choice between these two kinds of learners. The memory-hook panel has two parts, stacked, and both are always there: a "Your Memory Hook" box at the top that a student can type into at any moment, and the app's own "Suggested Memory Hook" underneath it, which stays hidden behind a Reveal button until the student asks for it.
That second part matters more than it looks. Answering the question does not reveal the prepared hook — a student who would rather invent their own association, or who doesn't want a canned mnemonic in their head at all, will never be shown one. The app's hook is genuinely opt-in, question by question, and it goes back to hidden every time a new state comes up. Writing your own is equally optional: the box can sit empty forever and nothing nags about it.
When the app's own "Suggested Memory Hook" is on screen, a small row of stars sits under it asking "Did this hook help?" If — and only if — a student taps one, the app sends a short message to a small program the author runs, and that message contains exactly two things:
That is the entire message. There is no name, because the app never asks for one. No session name — the "🧭 Golden Explorer" label stays on the device. No account, no login, no cookie, no device details, no location, no timestamp added by the app. And nothing a student wrote: the memory hook a student types in the top box is never sent, not even a fragment of it. The message itself contains nothing that identifies a student, a class, or a device — two ratings from the same student are indistinguishable from two ratings by two different students. (As with any web request, the server briefly sees the network address it came from, the same way any website's server does — that's just how the internet works, not something this app collects or stores on top of the state code and star count.) Each rating is kept for up to a year, purely so the author can see how ratings for a hook changed after he rewrote it, then it ages out automatically.
Why collect it at all. The memory hooks are the hardest part of this app to get right. A mnemonic that makes perfect sense to the adult writing it can be meaningless to a 13-year-old, and there is no way to tell which ones actually land except by asking the kids using them. The stars are that question. They show which prepared hooks are working so the ones that aren't can be rewritten.
What this doesn't change. Nothing else goes anywhere. Mastered states, streaks, the session name, the Kid/Classroom theme, the difficulty level and every hook a student writes themselves all still live only in that one browser, exactly as before. Rating is optional: skip it and the app behaves identically in every other respect. And if the device is offline or the message doesn't get through, the app just carries on — the rating is dropped and nothing breaks or waits.
When a student answers questions, the app remembers which states they've mastered so they can close the tab and pick up later. It remembers any memory hooks they wrote the same way. All of that is kept inside the web browser, on the one device they were using — the same place a website remembers that you prefer dark mode. It is not a record on the internet. There is no account and no login, and progress is never uploaded: the star rating described above is the only message the app ever sends, and it carries a state code and a star count, nothing about the student or their progress. If you open the app on a different tablet, that tablet starts from scratch, because the two devices have no way to know about each other.
That "remember where I left off" behavior is helpful for one student and confusing for a class. If a tablet gets handed from one student to the next, the second student would otherwise open the app and see the first student's mastered states already filled in green — and find the first student's hand-written memory hooks already sitting in the box.
So: each new student should tap "Reset Progress" before they start. It's the small button sitting directly above the map, on the same line as the "Mastered: 0 / 50" count. To avoid anyone wiping a student's work by accident, it takes two taps: the first turns the button red and it reads "Tap again to confirm", and the second tap does the reset. If nobody taps a second time within a few seconds, the button quietly goes back to normal and nothing happens.
After a reset, the map is empty again, the count is back to 0 / 50, every memory hook the previous student wrote is gone (their own writing is their own, not something the next kid should inherit pre-filled), and the app shows a new session name next to the count — something like "🧭 Golden Explorer" or "🧭 Prairie Ranger". That name is picked at random from a list of geography words. It's just a friendly, memorable way to see at a glance that this is a fresh, empty session and not someone else's leftover work. It is not a student ID, it isn't tied to any real name, and it doesn't identify or follow anybody — the app never asks who is using it, and never will. Two students who happen to get the same session name are simply two coin flips that landed the same way.
One reassurance worth stating plainly: resetting only clears the saved progress and hooks on that one device. It cannot delete or affect any other student's work, anywhere, because there is nowhere else — no record of any student's progress or writing exists at any point, on any computer but the one in front of you. And the star ratings are unaffected by a reset for the simple reason that they never contained anything belonging to a student in the first place: a state code and a number, with no way back to whoever tapped it.