Ask a student to search the internet for help with a Science chapter and you will get an answer within seconds — and that is precisely the problem. The answer might be correct, might be for a different curriculum entirely, might be three years out of date, or might simply be wrong in a way that sounds confident. We built the Library to be the opposite kind of resource: smaller, slower to grow, and answerable to a real name.
What the Library actually is
The Library is a shared shelf inside PDP Shikshya where staff — teachers and admins — upload or link study materials. In practice this covers a wide range: full syllabi, revision guides, sample tests written by the school's own teachers, and, notably, the actual grade 9-10 textbooks the school uses in class. Once something is in the Library, it can be shared with exactly the audience it is meant for — one student who needs extra support, a small group preparing for the same exam, an entire grade, a subject department, or the whole school.
This is a deliberately narrow kind of sharing. A teacher is not publishing to the internet; they are handing a specific document to a specific set of students, the way they might photocopy a handout and pass it down a row of desks. The platform just makes that handoff faster and keeps it organized in one place instead of scattered across messaging apps and printed sheets.
Vetted by a name, not by an algorithm
Every single item in the Library carries a visible tag: "Vetted by [teacher/admin name]." This is not a footnote — it is the point of the feature. When a student opens a revision guide, they are not looking at content ranked by a search engine's relevance score or an algorithm's guess at popularity. They are looking at something a named adult at their own school has looked at, decided is accurate, decided is appropriate for that grade, and put their name next to.
That accountability changes what the shelf can be trusted to hold. If a sample test turns out to have an error, there is a person to tell, not a support ticket into a void. If a study guide is excellent, the next teacher who wants to reuse it for their own section knows exactly whose judgment they are relying on. We think this matters more than volume. A library with two hundred vetted documents is more useful to a fifteen-year-old preparing for an exam than a search index with two million unfiltered results, because the two hundred were chosen by someone who understood the syllabus and the student.
A wider concern this responds to
This is not a uniquely Nepali problem, and it did not start with AI. For years, educators and librarians internationally have been raising a concern that gets more urgent every time the open web gets bigger: the sheer amount of available information keeps growing, but the amount of it that is reliable for a specific student, at a specific level, studying a specific curriculum, does not grow at the same pace. A search result can be technically true and still be the wrong country's syllabus, the wrong grade's depth, or simply outdated by a revised edition. Generative AI has added a newer version of the same worry — text that reads fluently and confidently regardless of whether it is actually correct for the question asked.
The response many schools and education systems have been exploring is not to abandon digital resources but to narrow them deliberately — building smaller, human-curated collections instead of relying on open search or open AI output as the first stop for a student doing homework. That is the same instinct behind the Library. It is one practical, local version of a much broader idea: that for young learners, a curated shelf checked by someone accountable is often worth more than an ocean of content checked by no one.
The Library and the AI tutor share a wall
The Library is not only a place for students to browse on their own — it also feeds directly into PDP Shikshya's AI Socratic tutor. Documents added to the Library become part of what we call the tutor's curriculum wall: the bounded set of grade-appropriate, syllabus-aligned material the tutor is allowed to draw on when it answers a question or guides a student toward one.
This is a large part of why the tutor does not behave like a general-purpose chatbot answering from the whole of the internet. When a student asks it a question, it is reasoning from documents a real teacher at that actual school vetted and uploaded — not from an unbounded, unaccountable pool of web content that may or may not match what that student is being examined on. The wall is a constraint we chose on purpose. A tutor that can answer anything, sourced from anywhere, is more impressive in a demo and less trustworthy in a classroom. A tutor that is scoped to what the school's own teachers put in the Library is less flashy, and considerably more accountable.
Why the growth is slow, on purpose
One honest trade-off worth naming: the Library will always be smaller than the open internet. It grows only as fast as teachers have time to upload and tag material, which means there will be gaps — a topic a student wants extra material on that simply is not there yet. We accept that trade-off deliberately. The alternative — opening the tutor and the resource shelf up to unfiltered search — would close those gaps instantly, but at the cost of the one thing the Library is actually for: knowing, for every document a student reads, exactly who stood behind it.
We would rather a student find fewer things and trust all of them, than find everything and have to guess which parts are right.
— PDP Shikshya product notes
What this means day to day
For a teacher, the Library is a way to make good material reusable — a revision guide built for one section this year can be shared with the whole grade next year without retyping anything. For a department head, it is a way to standardize what every section of a subject is working from, so that a sample test one teacher wrote reflects the same syllabus depth another teacher would expect. For a student, it means that when they go looking for help outside of class hours, the first and most reliable stop is a shelf built specifically for them, by people who already know what they are studying and why.
- Uploads and links are added by staff — teachers or admins — never by students
- Sharing can target a single student, a group, a grade, a department, or the whole school
- Every item shows a "Vetted by" tag naming the staff member responsible
- The same vetted documents form the AI tutor's curriculum wall, scoping its answers to real, school-approved material
None of this replaces a teacher's judgment in the moment — a student still asks their own teacher when something is genuinely unclear, and the tutor is built to guide rather than simply hand over answers. What the Library changes is what sits between the classroom and the rest of a student's studying: not a search bar that returns everything and guarantees nothing, but a shelf that returns less, and means it.