Search Multiple PDFs

Find one phrase across a whole folder of PDFs

The problem is always the same shape. Twelve bank statements, and you need the one with a specific transaction. A stack of contracts, and somewhere in them is the indemnity clause a colleague swears exists. Forty pages of lecture slides across six files, and the exam is on a term you can't place. Desktop search is unreliable with PDF content, and opening files one at a time to Ctrl+F through them is the kind of chore that eats an afternoon.

This tool does the loop for you. Upload up to 20 PDFs, type a word or phrase, and the server reads every page of every file and hands back an Excel spreadsheet of the results: one row per occurrence, with the file name, the page number, and a snippet of surrounding text so you can judge each match without opening anything.

How the search behaves

It's a plain text search, deliberately. Type invoice and it finds invoice, Invoice, and INVOICE, because the matching ignores case. Type a full phrase like net 30 days and it looks for that exact sequence. There are no regular expressions and no fuzzy matching, and that's a feature: you always know exactly what a match means, and a mistyped pattern can't silently match nothing while looking like it worked.

Every occurrence gets its own row, so a page that mentions your term three times produces three rows. The snippet shows roughly a sentence of context on either side, which is usually enough to tell the match you wanted from the coincidences.

The summary sheet keeps the batch honest

The second sheet in the workbook lists every file with its match count, including the zeros. That matters, because a zero has two possible meanings: the term genuinely isn't there, or the file couldn't be searched at all. Password-protected files and scans without a text layer are labelled as such in the summary, so a locked file never masquerades as a clean no.

For a password-protected file whose password you know, unlock it first and include the unlocked copy in your next batch.

That last case deserves a plain explanation. A scanned PDF is a photograph of a page. It looks like text, but there's nothing inside for a search to read until OCR adds a text layer. If your batch is scans, no search tool on earth will find anything in them as-is.

The fix is one step: run them through our OCR PDF tool first, then bring the searchable copies back here.

A real example

Say your accountant asks which month a particular supplier payment went out, and you have a year of statements as twelve PDFs. Upload all twelve, search the supplier's name, and thirty seconds later the spreadsheet says statement-march.pdf page 3, statement-august.pdf page 2, with the amounts sitting right in the snippets. Two payments, found, dated, documented, without opening a single file. The same trick works for finding a clause across contract versions, a name across meeting minutes, or an error code across log exports.

Once the results point you at the pages that matter, you can extract just those pages into a separate PDF if you need to share the evidence without the whole document.

Limits, honestly stated

• Up to 20 PDFs and 100 MB per batch, with the per-file limit shown on the upload zone. Bigger jobs work fine in batches.
• Search terms are 2 to 100 characters of plain text. No regex, no wildcards, by design.
• Results cap at 5,000 matches; if you hit that, the summary says so. A term that common usually wants narrowing anyway.
• Hyphenation and line breaks inside a phrase can hide a match, since the PDF's text is searched as the file stores it.

Privacy and file handling

Statements, contracts, and minutes are exactly the kinds of files people rightly hesitate to upload anywhere. The handling here is strict: encrypted transfer, processing by an isolated routine that reads text and writes a spreadsheet, and deletion of your uploads the moment the job completes. Nothing is stored, read by anyone, or reused. The results file you download is the only artifact, and it lives with you.