Your child already has an AI that will finish the worksheet. This is the one that won't.

Not by itself — and anybody who tells you their app blocks another app is selling you something. That switch lives on the device: Screen Time on an iPad or a Mac, Family Safety on Windows, Family Link on Android. But you can absolutely do it, it is free, and it takes about half an hour per device, so we wrote the instructions down in order with the screens named. The families who get the most out of Gnappo do exactly that: close the open search box and the general-purpose chatbots, and leave one door open that is age-banded, guarded and written down for you. The guide is here.

Four rungs, and you pick one per child. Questions and decomposition only, never stating the solution even when asked directly. Hints that get more concrete once your child has shown an attempt. The method worked through in full, then “now you try one”. Or the answer on request, for a parent who wants a reference. Hints is the default from about ten upwards; younger children start on the worked method, because a seven-year-old has to see a thing done before they can do it. You can change it whenever you like, and every change is written into your activity log with the value and the time — because “changed to Answers it at 21:40, the night before the exam” is a thing a household should be able to see.

No, at any rung, including the most permissive one. Gnappo never produces a finished piece of work to be handed in — not a complete essay, not a filled-in worksheet, not a whole lab report, not a paragraph written to be pasted. Ask it to write the essay and it will say, warmly, that it has to be in their words, and offer to help them get there. It will not work through a whole assignment in one go either, however many questions are pasted at once. The rungs decide how much of the thinking Gnappo does; this decides whose name is on the work, and it is not a setting.

Two things, and neither is a score. After it has helped, Gnappo asks one question back that only somebody who followed the explanation can answer — a different question, not the same one returned. And you can have it ask “what have you tried so far?” before it helps at all, which is the single most telling switch in the product: a child who cannot answer it has not read the question. Your evening summary then says something like “asked for the answer nine times, passed the check-back on two of seven” instead of leaving you to guess. Those are counts of what happened in a conversation. They are never turned into a mark.

It is not aligned to any particular state's standards or to your school's scheme of work, and we are not going to claim it is. Every question is written for a US grade level from 1 to 12 and for a named skill — fractions, photosynthesis, common denominators — so it lines up with roughly where a child of that age is, and Gnappo finds the right one from how your child has actually been answering rather than from what year group they are in. That last part is the useful difference: a child who is ahead in maths and behind in reading gets questions pitched separately in each, which a class of thirty cannot do.