BASALT · JOURNAL
How to search multiple PDFs on a Mac for the same term
Searching one PDF is easy; proving you searched a production consistently is harder. Files may have no text layer, use broken character maps, contain different spellings, or hide values in attachments and form fields.
Define the corpus
Create a manifest of the files and page counts you intend to search. Remove duplicates or record them deliberately. If a file is password-protected, resolve access before the search rather than silently skipping it.
Classify each PDF as born-digital, searchable scan, or image-only. OCR image-only pages locally and record which outputs were created. A folder search cannot find words in pixels.
Expand the term list
For names, include surname-first forms, initials, aliases, possessives, common OCR errors, and email usernames. For numbers, include punctuation and spacing variants. Search labels around an identifier when recognition may be poor.
Exact search is valuable because it is auditable, but add case-insensitive and pattern searches where appropriate. Review hits in context; a common surname or four-digit number can create many false positives.
Turn results into review items
Record file, page, matched term, surrounding context classification, reviewer decision, and reason. Do not copy sensitive content into an unrestricted spreadsheet. For redaction, marks should be created from approved hits and reviewed before application rather than deleting every match automatically.
Run a final search on the output corpus after redaction and sanitization. Also inspect attachments, metadata, and form values because ordinary page search may not include them.
Frequently asked questions
Why does Finder miss text in some PDFs?
The file may be an image-only scan, have a malformed or unsupported text mapping, or not yet be indexed. Use a PDF-aware search and OCR where needed.
Can I search handwriting across PDFs?
Handwriting recognition is less reliable than printed OCR. Use it as a suggestion and visually review relevant pages.
Should every search hit be redacted automatically?
No. Matches require context and an approved reason. Automation should build a review queue, not make disclosure decisions.
Doing it in Basalt
Basalt searches the whole open document, runs on-device OCR for scans, detects common PII patterns, and puts every suggestion into a mandatory review queue. Approved hits can be marked and verified without sending files elsewhere. Get Basalt.
Redaction that proves itself
Basalt destroys the content you mark, then re-opens the file it wrote and proves the content is gone before it saves anything. Your documents never leave your Mac.