Turning a pile of scans into documents you can actually find

• 300 DPI for text documents. It is the sweet spot for legibility and size, and it is what OCR engines are tuned for.
• Greyscale for printed text, colour only when colour carries meaning — a stamp, a signature, a chart.
• Feed in batches by document, not by pile, so you do not have to split as much later.
• Turn on deskew and auto-rotate if your scanner offers them; crooked pages hurt OCR accuracy badly.

A stack scan produces one long PDF containing many unrelated documents. Split it into individual files immediately, while you still remember what is in it. Extract the page ranges that belong together, and delete the blank backs and separator sheets in the same pass.

Use YYYY-MM-DD_party_subject. For example, 2026-04-12_northbank_mortgage-statement.pdf. Dates first means chronological sorting for free; lowercase and hyphens mean no problems across operating systems and cloud services.

Resist folder trees more than two levels deep. A shallow structure plus good filenames beats a deep hierarchy every time, because search works on names and nobody remembers which of eleven folders they chose.

Fix orientation before filing, not when you next need the document in a hurry. Rotating is instant and permanent, and a correctly oriented archive is far more pleasant to read on a phone.

• Tax and financial records: check your local requirement, commonly five to seven years.
• Warranties and receipts: keep for the life of the item plus a year.
• Contracts and property records: keep indefinitely.
• Follow 3-2-1 — three copies, two kinds of media, one off-site. A single cloud folder is one copy, not a backup.
• Test a restore once a year. An untested backup is a belief, not a backup.

• Split PDF
• Extract PDF Pages
• Rotate PDF