Abre mil páginas como una.

We measured that reader ourselves on a Mac, then built PageHawk around what we found. 30 preference panes behind seven tools became one pane. A Compare tool capped at exactly two files became one that takes a whole folder. A 1.68 GB update downloaded without asking became one that always asks first.

Every profession that reads for a living spends time the same way: a literature review screens hundreds of papers to keep dozens; a semester of readings runs to thousands of pages; a deal requires reading every draft of every document. Legal work is simply where that spend has been measured and published, so these four numbers come from there. Two carry a published source; the other two are an industry estimate and a planning assumption, each labeled as what it is:

Keeping the work on the device pays a second dividend, and it is the one most people underrate. Unpublished research, a manuscript, an unsigned agreement, a client file: pasting any of them into a consumer artificial-intelligence ("AI") chat tool hands a copy to somebody else's server. Where that has been tested in court the answer was unforgiving: a February 2026 ruling (United States v. Heppner, Southern District of New York, Judge Rakoff) held that a defendant's own exchanges with a public AI tool were not covered by attorney-client privilege, because the tool is not an attorney and no lawyer was party to them. PageHawk answers on your Mac, so whatever you are reading stays yours. And whether the hour is billed, grant-funded or simply your own, every quarter-hour the reader gives back goes to the work itself.

This is an interactive simulation of the reference-jump loop, shown here so you can feel it. Three independent tools (CoParse, Macro, Sioyek) each built an entire product around this one loop; PageHawk is building it into the reader itself, rather than into one more tool to switch to. The same loop reads a paper: citation to reference and back.

Triage 200 papers for a literature review, compare revisions, and surface the ones that actually matter.

Group related papers and methods, read figure-heavy preprints side by side, with nothing sent to a server.

A semester of readings and sources in one place: find the passage you half-remember, keep the sources that earn their citation, and export the list.

Compare filings, proposals, and deal drafts at the pace of the deal: quarter over quarter, version over version, in one pass.

Find the changed clause across a 500-document data room, and the near-duplicate pleadings buried in a discovery production.

Review scanned archives that need Optical Character Recognition ("OCR"): registers, manifests, and declassified files, read and matched on your Mac.

Every screenshot is the real interface, populated with genuine public records: court opinions, Securities and Exchange Commission ("SEC") and central-bank reports, arXiv papers, national-archive scans. In PageHawk they would be your own files, read and scored on your Mac.

The math: Acrobat Pro runs about $240 a year on its own. Add Adobe's AI Assistant, a separate paid add-on, and AI-inclusive Acrobat runs $300 to $340 a year. PageHawk at $129 once covers ask and compare, and Pro adds true redaction and the rest of the legal pack for $149 a year. Recovering 15 minutes a day is worth more than the reader costs at almost any hourly rate; at the $200 to $400 common in law and consulting it pays for a year of Pro in a single week. At those rates, the time lost to the reader costs more than the reader does.

Students & researchers: everything in Pro for $39, verified with your .edu email. Legal teams: the legal workflow (compare, true redaction, Bates) has its own page. The fairness bar: you renew only when you choose, canceling takes one click, there is no early-termination fee, and billing support answers. Acrobat Pro runs about $240 a year on its own, more like $300 to $340 a year with Adobe's AI Assistant add-on; at a $300 billable hour, one recovered hour covers nearly four years of PageHawk at $79 a year.