Can I search inside PDFs, photos and Office files?

Yes. Searching inside files is one of the main reasons Harbor exists.

Every PDF, photo, scan, and image you put in a note is indexed by server-side OCR, so ordinary searches reach inside them:

• The whiteboard photo from a planning session — search a phrase that was written on it.
• A folder of scanned receipts — search the store name or an amount.
• A contract PDF — search a clause, and Harbor highlights the exact line on the page.
• The photo of a serial-number sticker you took in the crawl space — search part of the number.

It works for PDFs with real text and for image-only PDFs (like most scans) — those get OCR’d the same as photos. And it covers attachments your device hasn’t downloaded yet, because the index lives on the server.

PDFs you save on a Mac with Save PDF to Harbor are searched the same way.

In search results, a file found by the words inside it is marked In file. That goes for every kind of file Harbor reads: photos, PDFs, the Office and text files below, and transcribed recordings.

Word, Excel, PowerPoint and text files

Harbor reads the words inside .docx, .xlsx and .pptx files, and their macro-enabled versions, .docm, .xlsm and .pptm. Search finds a word in one of them the same way it finds a word in a PDF. That includes a Word document’s headers, footers, footnotes and comments, a workbook’s sheets, and a deck’s slides with their speaker notes.

Plain text files work the same way: .txt, .md and .csv, plus .log, .text and .markdown.

None of this is OCR. These files already hold their words as text, so Harbor reads them straight out of the file, on our own servers. Reading them sends nothing to another company. Harbor goes back for the files already in your account too. Those take a while longer than new ones.

A few differences from a PDF:

• No highlight. There’s no picture of a page to draw on, so the match isn’t marked inside the file. Pick the result and you’re in the note that holds it.
• Spreadsheets hold values, not formatting. Harbor reads what each cell stores. A cell showing $12,500.00 stores 12500, so search for that. Excel stores most dates as numbers, so a date in a cell usually won’t be found.
• The first megabyte of text. Harbor reads up to the first megabyte of text in each file. That’s a few hundred pages of writing, but a big spreadsheet, CSV or log can run past it. Sheets are read in tab order, and nothing after the limit is searched.

Some files Harbor can’t read inside:

• Old .doc, .xls and .ppt files, the formats Office used before 2007.
• Files that need a password to open.
• OpenDocument files (.odt, .ods, .odp), and Apple’s Pages, Numbers and Keynote files.

Save a copy as a PDF, or as a modern Office file without a password, and Harbor can read that. Other kinds of files, like RTF or HTML, are found by their name only. A file’s name is searchable either way.

File names

Search reads file names too, extension included. budget finds Q3 Budget.xlsx, and xlsx finds your Excel files. How file names match.

Two search-grammar helpers pair well with this. The resource: operator narrows results to notes carrying a certain kind of attachment, and quoting a phrase ("grand total") keeps matches exact. The full grammar is in What search operators does Harbor support?

The one standing exception: encrypted notes and their attachments are never read or searched, file names included — Harbor truly can’t read them, which is the point of encrypting.