Paste it.

Whatever an AI produced — MODU takes it in without breaking it, lets you tidy it, and shelves it. Math that used to shatter on paste becomes a question; questions become an exam paper.

01 Take in · 02 Tidy · 03 Shelve · 04 Print

Open the studioSign in with a Google account.

Paste something first, no sign-in ↓

Source — click another question to switch

Here are four exam-style problems. The two geometry ones come with figures.

1. Let $F$ be the focus of the parabola $y^2 = 4x$, and let $H$ be the foot of the perpendicular from a point $P$ on the parabola to its directrix. If $\overline{PF} = 5$, what is the area of triangle $PFH$? [4 pts]

```figure w=58%
⎔ 도형 장면 JSON 836자
```

① $8$
② $9$
③ $10$
④ $12$
⑤ $15$

2. In the coordinate plane, let $F$ and $F'$ be the foci of the ellipse $\frac{x^2}{9}+\frac{y^2}{4}=1$. A point $P$ on the ellipse satisfies $\overline{PF} = 2\overline{PF'}$. What is the area of triangle $PFF'$? [4 pts]
① $2$
② $3$
③ $4$
④ $2\sqrt{5}$
⑤ $6$

3. A cubic function $f(x)$ with leading coefficient 1 satisfies the following conditions.

> (a) $f(0) = 0$, $f'(0) = 0$
> (b) The function $|f(x)|$ fails to be differentiable only at $x = 2$.

Find the value of $f(4)$. [4 pts]

4. For three distinct points $O$, $A$, $B$ in the plane, $|\overrightarrow{OA}| = 2$, $|\overrightarrow{OB}| = 3$ and $\overrightarrow{OA}\cdot\overrightarrow{OB} = 3$. What is $|\overrightarrow{OA}+\overrightarrow{OB}|$? [3 pts]
① $\sqrt{13}$
② $4$
③ $\sqrt{19}$
④ $2\sqrt{5}$
⑤ $\sqrt{21}$

조판

Let $\displaystyle F$ be the focus of the parabola $\displaystyle y^2 =\nobreak 4x$ , and let $\displaystyle H$ be the foot of the perpendicular from a point $\displaystyle P$ on the parabola to its directrix. If $\displaystyle \overline{PF} =\nobreak 5$ , what is the area of triangle $\displaystyle PFH$ ? [4 pts]

\displaystyle y^2=\nobreak 4x

\displaystyle H

\displaystyle O

\displaystyle F

\displaystyle P

\displaystyle y

\displaystyle x

①
$\displaystyle 8$
②
$\displaystyle 9$
③
$\displaystyle 10$
④
$\displaystyle 12$
⑤
$\displaystyle 15$

AnswerDifficultyConceptSource

Drag the red point P. It stays on the parabola, and the foot H follows.

01 · Take in
Paste, drop a file, or one photo.
Paste a ChatGPT, Claude or Gemini answer whole. Math with no delimiters, absolute-value bars inside a table, an unclosed $$ — all of it comes in as is. The raw text stays on the left; the typeset document sits on the right. Glyphs that could not be recovered are counted and marked, never invented.
Pasted text
The AI answer as written. Broken math and tables are recovered.
.hwpx
Equations, tables and figures read whole; questions split on endnote boundaries.
One photo
A photographed problem is read into a question. Several pages? Use a PDF.
PDF
Question extraction from PDF is being prepared with vision reading.
Raw paste (left) and the typeset document (right). Lost glyphs: 0 — and 13 □ marked as already missing in the source.
Intake — what arrived, how many questions it looks like, what needs a hand. Three sentences.
02 · Tidy
You draw the boundaries. Then one question at a time.
Drag across the raw text and that span becomes a question. Wrong grab? Remove, merge, undo. Then the workbench, one question at a time — raw on the left, typeset on the right. Promote math with ⌘M, fill in answer, difficulty, concept and source, and put it on the shelf.
Boundaries — three questions grabbed by dragging. Faded text is not a question yet.
The workbench — raw and typeset side by side. One question, with answer, difficulty, concept and source.
02 · Figures
A figure in a photo becomes a shape you can edit.
Figures on exam papers arrive as photos. 'Convert to editable' reads the relations in the photo — focus, directrix, intersections — rebuilds the figure from them, and fits the draft's free numbers to the ink in the photo. After confirming, you can still drag the points.
Review the conversion
Later (esc)DiscardConfirm as figure block
Original photo — kept as is
Converted figure — still editable after confirming
$\displaystyle (y+\nobreak 2a)^2=4p\mkern1.5mu(x-\nobreak a)$
$\displaystyle y^2=\nobreak 4px$
$\displaystyle x=\nobreak -p$
$\displaystyle D$
$\displaystyle B$
$\displaystyle E$
$\displaystyle O$
$\displaystyle C$
$\displaystyle G$
$\displaystyle F$
$\displaystyle P$
$\displaystyle A$
$\displaystyle y$
$\displaystyle x$
drag the red point P — the chord turns and the intersections and feet follow
03 · Shelve
One star, one question.
Shelved questions appear as stars on the first screen. Concepts are hubs, questions are stars — constellations from afar, single questions up close. The library only unfolds once you have an intent: a search term, a category, a source you just imported.
The first screen — the library as constellations. The paste box sits below.
The library — category index, recent imports, recent questions.
04 · Print
Scope it, order it, A4.
Pick a scope by source and concept and it becomes an exam paper. Adjust order and points, add school and exam names, and check A4 pages in the preview. Save as PDF from the print dialog and you have a file. Preview and print are the same rendered screen, so they cannot differ. What ships today is the question sheet — answer and solution sheets are not there yet.
Exam paper — layout settings and the A4 preview. KoPub Batang, the look of Korean CSAT papers.

No sign-in. Right here.

The same engine the studio uses, running in your browser. Real output from ChatGPT, Claude or Gemini — or one of the cases below. It typesets as you type.

Fifty thousand characters. One pass. Zero errors.

No headings, no bold — Korean and English prose with hundreds of formulas, block equations jammed against paragraphs on purpose. Your browser typeset the whole document just now. Scroll it.

This page was typeset by the engine too.

The before/after above and the playground are produced by the same code the studio runs. The engine depends on katex and jszip and nothing else, so it behaves the same at build time on a server and live in a browser. It is not published separately yet — it is hardening inside the studio first.

Where things stand

Under construction. Screens change often.
One account type: Google.
Question extraction from PDF is in preparation.
Answer and solution sheets are not printable yet.

Paste it.

Paste, drop a file, or one photo.

You draw the boundaries. Then one question at a time.

A figure in a photo becomes a shape you can edit.

One star, one question.

Scope it, order it, A4.

No sign-in. Right here.

Fifty thousand characters. One pass. Zero errors.

This page was typeset by the engine too.

Where things stand