📐 For engineering students

An AI study tool for engineering students that knows its limits

Engineering exams lose marks in two places: the algebra, and the concept the algebra was expressing. This tool is useless at the first and genuinely good at the second.

Questions per set

Bigger sets read more of your document. Included with Student and Genius.

PDF, PowerPoint, Word, or a link · up to 10 MB · start without an account

What an engineering course actually gives you

  • Lecture notes and theory handouts

    The written explanation around the equations, which is where the assumptions, the boundary conditions, and the reasons live.

  • Lab briefs and reports

    The method, the sources of error, and the discussion. Heavily assessed and almost never revised because nobody makes cards from a lab brief.

  • Standards and design guides

    Extracts from codes of practice read cleanly as PDFs and are exactly the reference material a design module examines you on.

📐 See the output firstIllustrative sample

What does an engineering card look like?

A passage from a structures lecture, and the card generated from it. Sample material, written by us to show the shape of the output.

Your lecture notes

Structural analysis, lecture 9 notes.pdf

Bending stress varies linearly with distance from the neutral axis, so material near the axis carries almost none. This is why an I-beam concentrates area in the flanges: the same mass placed further out raises the second moment of area.

The card it writes

Why does moving material away from the neutral axis make a beam stiffer in bending?

Second moment of area depends on the square of the distance from the axis, so the same mass placed further out contributes far more resistance to bending.

The card asks why the I-beam shape works rather than for the formula, so answering it needs the relationship between distance and stiffness rather than a memorised expression.

The scored quiz runs on every plan including Free. Flashcard mode, shown here, comes with Student and Genius.

The same three steps, whatever you study

  1. 1

    Upload what your course gave you

    A PDF, a lecture deck, a Word document, a pasted link, or text typed straight into the box.

  2. 2

    It writes questions from that file

    Not from a general model of your subject. Only from the document in front of it, which is what makes the questions match your module.

  3. 3

    Answer, miss some, come back

    Questions you get wrong join a weak pile and return in the daily review. Mastery lapses after a fortnight, so nothing stays finished forever.

What it will not do for engineering

  • It does not do maths. It will not solve, check, or set a numerical problem, and any number produced by a language model should be treated as unverified.
  • Derivations are not walked through step by step. It can ask what an assumption was for, but not why line four follows from line three.
  • Circuit diagrams, free-body diagrams, and plots carry no readable text, so a handout that is mostly figures produces very little.

Engineering student questions

Can it help with problem sheets?

Not with the working, no. It does not solve problems, check your arithmetic, or generate numerical questions with worked solutions, and you should be wary of any tool that claims to. There is a documented failure mode where language models lose accuracy when only the numeric values in a question change, and a problem sheet is precisely that pattern repeated with different numbers. What it does help with is the layer people skip: the concepts the problems are testing. Most marks lost on an engineering paper are not lost in the algebra but in choosing the wrong approach, missing an assumption, or misreading which model applies. Questions generated from your lecture notes attack exactly that, and they are quick enough to run before a problem sheet rather than instead of one.

Which parts of an engineering degree does it actually suit?

The text-heavy parts, which are a bigger share of most degrees than students expect. Materials, thermodynamics theory, structural concepts, control theory in words, engineering ethics, sustainability, project management, and every lab brief you have ever been assessed on are all substantially written rather than numerical. Those generate excellent questions. The purely computational modules generate poor ones, because the examinable content is a procedure rather than a set of propositions, and a question about a procedure written in prose does not test whether you can execute it. A useful rule of thumb is to look at a handout and ask whether a well-informed person could be examined on it in an oral. If yes, it will produce good questions. If the page is mostly equations and worked examples, it will not.

What about lab reports and coursework?

Lab briefs are one of the highest-yield things you can upload and almost nobody does it. They are dense, entirely text-based, heavily weighted in most modules, and they contain the exact material that gets examined in vivas and short-answer sections: why this method, what the dominant source of error is, what assumption the analysis rests on, and what the result would look like if it failed. Nobody makes flashcards from a lab brief, which means almost nobody revises them, which means the marks are sitting there. Upload the brief before the assessment and answer questions on your own method. What it cannot do is write, mark, or improve a report you have drafted, since there is no long-form feedback of any kind here. That half stays between you and your demonstrator.

Comparing tools first? Read our roundup of AI study tools for engineering students.

The lab brief nobody revises

Free to start, no card required 🥰

Turn it into questions