A PDF Reader for Research, Not Just Viewing

Every device ships with a PDF viewer, so if you're searching for a PDF reader, a viewer isn't what you're missing. What's missing is the gap between displaying a document and reading it — a gap that gets wider the longer and denser the document is. For research reading — papers, reports, theses, technical books — that gap is the whole problem.

What a viewer can't give you

A viewer shows you pages. Research reading needs the document's structure: where the sections are, how far through you are, what's a footnote versus body text, which words are the author's and which are the page's. PDFs encode almost none of that — they're printing instructions, not text — so any app that wants to offer it has to reconstruct it.

That line — headers and page numbers gone, prose in reading order — sounds small and changes everything downstream. Velo's on-device import pipeline goes further: sections and chapters detected into a hierarchical table of contents with per-chapter progress, two-column layouts resequenced, footnotes and endnotes extracted and linked to their reference points, bibliography and appendices recognised and separated, OCR artifacts cleaned. Scanned documents go through Vision OCR on the device, with an optional credit-metered server pass for the genuinely hard cases — invoked by you, not silently.

The output is a document that behaves like a book: navigable, measurable, readable. On an iPad — the natural surface for this work — that's the difference between reading a paper and squinting at a photograph of one. (The full workflow is in reading academic PDFs on an iPad.)

The reading part

Structure gets you to clean prose; the remaining problem is the one research readers know best — attention across forty dense pages. Velo gives you three ways through the same text, sharing one position:

• Reader mode, the typeset page, for figures, equations, and passages that need your eyes free.
• RSVP mode, the text streamed word-by-word at a fixed focus point at a pace you set — a focus lock that makes a methods section impossible to skim by accident. (The honest research on RSVP, limits included.)
• Scrub, a momentum drag back through the stream for the definition that didn't land.

Because the position is shared, switching is free — stream the prose, page-read the figures, and never hunt for your place.

For triage, the on-device intelligence dashboard summarises a single section or chapter — more useful than whole-document summaries, which mostly re-derive the abstract — and answers questions scoped to the document. All of it runs via Apple Intelligence on your hardware, which matters when the PDF is an unpublished manuscript or an internal report: it never leaves the device.

Practicalities

Import from Files, Safari, or Mail via the Share Sheet; documents store locally and read fully offline. Both reading modes, import for 3 documents, stats, and the dashboard are free — enough to run your three ugliest PDFs through the pipeline and judge the output yourself. Pro removes the cap and syncs your position across iPad, iPhone, and Mac, so the paper you triaged on the phone is waiting at the same paragraph on the bigger screen.

← All guides