Frequently asked questions.

It takes the two documents every course already gives you, an outline and a timetable, and turns them into a working semester: a calendar with your classes and study sessions on it, an assessment tracker with checkpoints, practice papers, and a spaced-repetition deck built from your own material. Flux, the built-in tutor, answers questions about the work and computes anything computable with a deterministic engine rather than guessing at it.

Any of them, anywhere in the world. Course outline import takes a PDF or a URL from any institution, and timetable import takes a standard .ics file or a photo of a printed grid. None of it is university-specific, and none of it cares whether your LMS is Canvas, Blackboard, Moodle or Brightspace. Resource mirroring, the desktop-only feature that copies your course files into a tidy folder tree, works with Blackboard and Canvas; Moodle and Brightspace mirroring isn’t wired up yet, though import works regardless of which LMS you use.

Both. The data model is a year → semester → course tree, which fits either, and the timetable-from-a-photo importer exists precisely because a high school student has a printed grid rather than an iCal feed. The tutor covers science, technology, engineering and maths at school and early-undergraduate level.

Very few. It is in testing, and that is the honest answer. There are no student numbers on this site because there are none worth quoting yet. What you get instead is that every feature described here is built and running, and that feedback from an early account actually changes the product.

Nothing, while it is in testing. There is no paid plan to buy, no card to enter, and no credit balance to top up. When pricing does arrive you will be told before anything changes.

Not on the free plan. You can type the question, paste it, or attach a PDF, all of which get the same verified treatment. Photographing a problem and asking about it is one of the three things held for a paid plan later, because reading an image costs several times what reading text does and a tutoring session is dozens of turns.

Not on the free plan. Generating one is a dozen separate model passes: a plan, an author and an examiner for every question. A single paper costs about as much as a whole week of free AI capacity. Offering it would really mean offering a week of the tutor or one paper, which is not a choice worth putting in front of anyone. The tutor, recall and your own files are unaffected.

Yes, and that is deliberate. Importing a timetable from a photo or screenshot is included on the free plan: it is a one-off setup step, not a per-question cost. The distinction is where the image is used, not what kind of file it is.

There is a daily fair-use budget per account. It exists so one heavy user can’t exhaust the shared capacity in the middle of everyone else’s exam week, not to push you toward an upgrade. Ordinary use does not come near it.

That the number in the answer was produced by a deterministic engine, the same arithmetic every time with no model involved, and that the engine’s result is what the explanation was written around. It covers the value, not the prose: a verified answer can still be explained clumsily, and the badge is not a claim about the wording.

No badge appears. Proofs, ε–δ limits, convergence arguments and anything with a symbolic parameter in it are not verifiable by an engine, so they are answered as tutoring with no guarantee attached. For a course that is mostly proof work, treat Flux as a tutor rather than a checker. That is the honest description of what it is doing.

It isn’t built for that and it doesn’t try to hide the fact that submitting AI text is misconduct at every university that has a policy on it. What it is built for is the part that is legitimately yours: planning the work, practising under exam conditions, and finding out which parts you cannot yet do.

Only if you turn that on, and it is off by default. Building recall cards from your notes and answering from your own course material both require an explicit, revocable consent, and revoking it deletes the index rather than just hiding it. Settings shows exactly what has been read.

In Australia. That covers both the database and the app, so your coursework never leaves the country to be stored. The specifics are in the terms, under Legal.

No. Accounts are isolated at the database level by Postgres row-level security, not only by application code. A query for your data cannot return somebody else’s even if the code above it has a bug.

No third-party analytics script, no session recording, no advertising pixel. Sign-up counts and AI spend are recorded first-party, in the same Australian database, because shipping usage data offshore would contradict the answer above.

Yes, for macOS. It holds a local-first copy of your data and syncs when it can, so a dropped connection during study week costs you nothing. Resource mirroring lives only there, because it needs a real browser and a real filesystem, which a serverless host does not have.

Yes. Blackboard (Ultra) and Canvas both work: the desktop app mirrors your course files into a tidy, week-first folder tree and pairs each problem sheet with its solutions. In both cases you sign in yourself, in a real browser window, including your university’s multi-factor step; only the resulting session is kept, encrypted, on your own machine, and your username and password are never seen, sent or stored.

Yes. The same sign-in, the same data, synced both ways. You can plan on a laptop and revise on the desktop app without exporting anything.