What happens between “go practice” and a better score.

A practice session warms up: it opens easy and climbs question by question to the difficulty your level earns, whatever the answers were. Kids pick Easy, Normal or Hard each session (a cat, a dog or a dragon): Easy aims a notch easier, Normal at their level, Hard above it. Mock tests start at that difficulty and move one step up on a correct answer, one step down on a miss — the same rule the digital SHSAT uses. A quarter of picks stretch one step above target so harder questions never come as a surprise on test day.

Reading questions come as passage sets: read once, answer the set, revise within it, then submit — exactly the real navigation rules.

• Train — level-scaled length, per-question clocks optional.
• Simulate — a full 50-question section on one 90-minute clock. A two-section 180-minute test with the break arrives in October.
• Speed Drill — ten minutes of arithmetic, as many as you can. Mental math is half the math section.

Multiple choice, numeric entry, multi-select, reorder, inline dropdown, drag-to-category, hot text, and point graph — every format on the digital test, so no seconds are lost to unfamiliar mechanics. Hunter and SAT writing add free-response prompts.

A wrong answer opens a short teaching moment — the concept, a worked example, the technique — and a retry before Next unlocks. Kids do not get to click past a mistake.

Everything missed goes on a review ladder (1, 3, 7, 21, 60 days). One in every three questions in a normal session is a review item that is due, so weak spots come back until they stop being weak.

Every question ever missed becomes a flip card: prompt on the front, answer, explanation and technique on the back. Cards flip to “learned” when the latest attempt is right.

A logistic ability fit over recent answers, simulated onto a 50-question section and mapped to the scaled score. Unlocks after 15 answers per section, shown with the published cutoffs for the specialized high schools. Hunter shows a readiness band; SAT a 400–1600 estimate.

Parents set a target and a test date. A nightly snapshot of the estimate builds a trend; once three weeks exist, the dashboard shows where the current pace lands on test day — as a range, never a promise — and emails if it drifts off track for two weeks running.

Accuracy per topic, weakest first, on the kid's page and the parent's. Questions are drawn evenly across topics, and a parent can boost any topic's coins to nudge a weak one.

Hunter and SAT writing prompts carry a rubric. Submissions are graded against it on a schedule — scores per criterion, two strengths, two concrete fixes — with coins scaled to the score. There is no chat: a kid writes, submits, and sees feedback next time they open the app.

Each essay is safety-checked, then comes to the Parent Zone first. One tap shows the AI's score and notes; the parent unticks any they don't want, adds a note and sends it back. Anything off-topic or concerning isn't scored: most go to the parent with a reason, and a few safety concerns go to our team first (privacy policy). The kid never sees a flag.