The ones worth asking first.

No, and we are not going to imply otherwise. What exists today is the design, the database — tenancy, capability grants, the append-only ledgers, the sync and conflict rules — and 674 hand-authored curriculum nodes across reading, spelling, grammar, mathematics, citizenship, science, nature study, geography, history, life skills, physical development and the arts.

What does not exist yet is the mobile app, the parent portal, and the generation pipeline in a form you could install. Beta families get builds as they become usable, starting with the parts that work on their own.

A place in the private beta, and an early role in building the place library. We will write to you when there is something to try, not on a schedule. If that turns out to be quiet for a couple of months, it will be quiet for a couple of months.

There is no automatic confirmation email, because there is no mailing list behind this yet. Replies come from a person, by hand — which works at a hundred families and is a good part of why we are only taking about a hundred. If you asked a question on the form you will get a real answer to it.

A family of seven — two parents and five children — living and homeschooling full-time in an RV. This started because nothing on the market fit, and it is being built to be used every day by the people building it. That is the main quality control mechanism and it is a good one.

Because they are. We would rather ship a site that says “this film is in production” than stage footage of an app that does not run yet. As real captures and real film exist, the placeholders are replaced. None of the screens described on this site are invented — they are all in the build plan.

That is the central design constraint, not a feature. Lessons, maps, field guides, capture and the whole daily loop run with zero bars. Vector basemaps are read straight off local storage. If a feature in the daily loop needs a network, we treat it as a defect. The only things that genuinely need signal are looking up a place’s current hours and opening a map application for driving directions — both of which you do while planning, when you have signal anyway.

Yes — see Adventure Mode. Same engine, no curriculum scaffolding, aimed at a family trip rather than a school year.

As many as you have. Pricing is per family and never per child, because charging per child punishes exactly the families this was built for. Children on a milestone-only track — infants and toddlers — get full profiles and full participation with no lesson scheduling, and switching one to the academic track later is a setting, not a migration.

Because a two-year-old should not be handed a device, and building him a screen would have been the easy thing to do. He gets tag-along mode inside the parent app: observation prompts an adult scores, producing a real record with his name on it and no screen time attached. This is a deliberate position, not an omission.

No. RV LIFE Trip Wizard has no public interface, so we bridge through the files it already exports — drag a GPX or CSV into the portal, or share it from your phone. Re-importing the same trip after you change it shows you a diff rather than duplicating everything, because you will re-export the same trip five times. We export GPX back out too. The goal is to augment your planner, not compete with it.

Nothing during the beta, and no card is collected. Beyond that we will be honest that the numbers are still hypotheses: a family plan in the region of twenty to thirty dollars a month for the full curriculum tier, with Adventure priced substantially lower and possibly per-trip. Annual at a discount. A full month of free trial, because a curriculum decision cannot be made in seven days.

Beta families who contribute places and lessons keep founder pricing for as long as they keep an account. That is not a launch promotion; it is the only fair way to treat the people who made the library worth having.

No. Children have no social surface — nothing about a child is reachable outside your household without a deliberate, per-item, revocable act by a parent. This is enforced by database row-level security with an automated test suite that runs on every schema change, not by a user interface remembering to filter something. There is no directory, no profile discovery, and no feed.

Not now, and when there eventually is a sharing layer it will start as text lesson plans only — no photographs, no messaging, no profiles beyond a display name. Each step after that has a gate in front of it. We would rather move slowly here than explain later why we did not.

No. Child references are pseudonymised before anything leaves the platform for lesson generation, and nothing about your children is used to train or develop models. Under the amended COPPA rule, disclosing a child’s personal information to a third party for AI training requires separate verifiable parental consent; we avoid the question entirely by not doing it.

The annual archive export ships in version one: full structured data, rendered PDFs, a media archive and a self-verifying manifest, written to storage you control. A twelve-year record outlives most software, and designing for our own disappearance is part of taking that horizon seriously.

It will keep you documented, which is a stronger position than most requirements ask for. Requirements vary enormously by state — Texas, where this was built, requires a bona fide curriculum in visual form covering five named subjects and essentially nothing else; other states require attendance logs, notification, testing or portfolio review. The records system was built for a future admissions officer rather than for a state, which is the harder bar. But nothing on this site is legal advice, and you should check your own state’s rules.