PyQDeck started as a simple question: why is finding last year's exam paper for a specific BEU subject so much harder than it should be? Everything below is a feature that exists because it solved a real version of that problem.
Browsing by semester and subject
The home page is organized the way the syllabus actually is - by semester, then subject. Each subject page groups its questions by year and by chapter, so you can either read through a full paper as it was originally set, or pull every question from one chapter across every year it's appeared, side by side.
Two kinds of search, and why both exist
Keyword search finds questions that contain the exact words you typed - fast and predictable, the right tool when you already know roughly what you're looking for.
Semantic search works differently: it understands the meaning of your query, not just the words in it. Search "how connections get set up between computers" and it can surface a TCP handshake question even though none of those words appear in it. This matters most when you remember the concept an exam asked about but not the exact phrasing - which, realistically, is most of the time.
Search results actually combine both automatically, ranking a question higher when it shows up as a strong match in either method - so you get the precision of keyword matching and the flexibility of semantic search in one search box, without having to pick.
Repeated questions
Every question page checks whether that question (or a close rephrasing of it) has appeared in a different year's paper, and shows you the years it repeated in if so. This is the single fastest way to find high-value questions - a question that's been asked three times across five years is a very different priority than one that's only ever appeared once.
Worked solutions, with voting
Where a solution exists, it's shown directly on the question page. Solutions carry a simple thumbs up/down so the community can flag ones that are wrong or unclear - no login required, one vote per device.
Filtering a subject's questions
Inside a subject, you can narrow the question list down by exam year, by chapter, or both - useful when you want to study one chapter's history across every year at once, rather than reading full papers front to back.
Connecting an AI tool
PyQDeck also exposes its question archive as an MCP server - which means you can connect it directly to Claude, or any other MCP-compatible AI tool, and ask it questions about your syllabus directly, with it pulling real questions and solutions from the archive rather than guessing. If you already use an AI assistant to study, this lets it work from your actual exam history instead of generic answers.
No account needed
Nothing on PyQDeck requires signing up. Browsing, searching, and voting on solutions all work anonymously - the archive is meant to be a reference you can open and use immediately, not a product you have to onboard into.