A PDF records positions: this line begins here, that heading sits there, and the footer occupies the bottom margin. Reflow changes the reading surface without changing the source file.
What reflow does
Paper reads the selectable text in a public PDF, identifies useful blocks such as headings and paragraphs, and presents them in a responsive column. On a phone, the text can use the full width available. On a desktop, the line length can stay measured instead of stretching across the whole window.
Adapts to width
Text wraps to the viewport rather than remaining trapped inside a miniature printed page.
Restores rhythm
Paragraph spacing, headings, type size, and line height support sustained reading instead of page navigation.
Clears repetition
Repeated page numbers and running headers can be removed when the extraction heuristics identify them safely.
Leaves source context
Reflow is a reading layer, not a replacement for the original PDF. Keep the source for exact page references and visual checks.
What reflow cannot solve
A scanned PDF has no text layer for the current reader to extract. Tables, equations, columns, charts, forms, and layouts where position carries meaning can also become imperfect when moved into a single reading stream. Those are not bugs to hide with confident copy; they are reasons to keep the original view nearby.
Currently supported formats
Paper's reflow reader currently accepts public HTTP(S) links to selectable-text PDFs up to 30 MB. OCR and non-PDF formats are future directions, not current capabilities.
See current support and future directions
Try reflow on a public PDF and judge the reading surface for yourself.
Open a PDFQuestions
Is PDF reflow the same as converting a PDF?
Not exactly. Paper creates a browser reading view from the PDF's text; it does not rewrite or export the original file.
Does reflow work on every PDF?
No. Selectable, text-first PDFs are the best fit. Scans and visually complex documents may not reflow usefully.