Skip to content

PDF converter prompt

Version 1.0.0

Test-series solutions and coaching notes arrive as PDFs. This prompt converts them, page by page, into Laya Copy format — one block per question, each with atomic bricks — so a whole booklet lands in Laya's import inbox for review. It works with any AI that can read a PDF; on a paid pass Laya can also do this in-app.

How to use

  1. Copy the prompt below.
  2. Paste it into your AI and attach the PDF (or paste the text of a few pages at a time).
  3. Read the blocks it returns; fix anything it flagged as unclear.
  4. In Laya, open Copy → Paste, paste the blocks; more than one question lands in the import inbox for review.

The prompt

Laya · Universal PDF converter prompt · v1.0.0

I will give you a PDF (or pasted pages from one): a test-series solution booklet, a coaching module, or my own notes, containing exam-style questions with answers. Convert its contents into the Laya Copy format below — one block per question — so I can paste the whole thing into Laya's import inbox and review it there.

Work through the document in order. For EACH question you find:

1. Take the question text as printed. If the document prints the paper, year, or marks near it, capture them in the tag; if a marks figure appears with the question ("(10)", "10 marks", "15M"), it is the marks.
2. Take the model answer as printed, verbatim, keeping its paragraphs. Do not rewrite it, shorten it, or fix its grammar. If the answer runs across pages, join it. Skip evaluator remarks, marking-scheme notes, and page furniture (headers, footers, page numbers, watermarks).
3. Extract BRICKS from that answer: the atomic facts, examples, data points, and arguments a person would need to recall to rebuild the answer. One thing per brick, self-contained, specific, under 200 characters (never over 280), and ONLY what the answer contains — never your own additions. Star (mark with an asterisk) the bricks a marker would look for first: the direct answers to the question's dimensions and the specific examples or data that prove them; at most a third of the bricks. At most 30 bricks per question — keep the most examiner-relevant if there are more.

THE FORMAT for each question — exactly this structure:

Q: <the question, on one line> [<paper>, <year>, <marks> marks]
A: <the answer as printed, keeping its paragraphs>
Bricks:
- <a brick>
- * <a starred brick>

Rules for the output:
- "Q:" then one space starts each block. The square-bracket tag is optional; include the parts you know (paper, year, marks — in that order, comma-separated); omit the brackets if you know none.
- "A:" then one space, then the answer; further answer lines follow without any prefix.
- "Bricks:" alone on its own line.
- Each brick starts with "- " (hyphen, space); a starred brick with "- * " (hyphen, space, asterisk, space).
- One blank line between blocks. Nothing else — no headings, no numbering, no commentary between blocks.
- Keep the document's order. If two questions share one answer, output two blocks with the same answer.
- If a passage is unreadable or a question has no answer in the document, output the line "UNCLEAR: <what and where>" instead of a block for it, and continue.
- If the document is long, work a few pages at a time and tell me where you stopped ("STOPPED AT: page N") so I can say "continue".

Begin with the first question. Output only the blocks (and any UNCLEAR / STOPPED AT lines).