How to use the pdf text search
- Choose a PDF, up to 25 MB and 500 pages. The tool reads the text of every page on your device and shows a progress bar. You can cancel while it reads.
- Type a word or phrase in Search for, then press Search. Tick Match upper and lower case or Whole words only if you need them. The results update when you change an option.
- Read the summary of how many matches there are and on how many pages, then the list of pages with a few lines of context around each match.
- Press Copy results to copy the page numbers, counts and snippets as text.
A word on two pages of a 10 page PDF
Searching a 10 page PDF for a word that appears once on page 2 and once on page 7 returns 2 matching pages and 2 matches, with a snippet from each. The summary reads "2 matches on 2 of 10 pages".
Options change what counts. In a page that reads "Invoices and INVOICE numbers; invoice-2 and re-invoice", searching for invoice with case ignored finds 4 matches, with whole words only it finds 3 (INVOICE, invoice-2 and re-invoice, because a hyphen separates words), and searching for Invoice with case matching and whole words only finds none on that page. Counts for each option were checked against a separate regular expression search.
How the search works
The tool uses the same reader a browser uses to show PDFs to pull the text from each page, then searches that text. Line breaks and repeated spaces are treated as single spaces, so a phrase that wraps onto the next line is still found. Searching is plain text matching, not a regular expression, so characters such as a dot or a dollar sign mean themselves.
Each page shows up to five snippets with the match highlighted, while the count includes every match. Matches do not overlap, so searching for aa in aaaa finds two. Snippets are cut with ... where the text continues.
When a PDF has no searchable text
Only text that is stored as text can be found. A scan or a photo of a page holds a picture of text, and this tool does not read pictures. If no page has any text, the tool says so. In that case, a scanner program with character recognition would be needed to make the text searchable.
Some PDFs store text in an odd order or break words into pieces, so a word may be split in the file and not found by an exact search. If a search misses something you can see, try a shorter part of the word. To read all of the text at once use PDF to text, to compare two versions try compare PDF, and to see the structure use the PDF outline viewer. Your file stays on your device.
Assumptions and limits
- Only PDF files are accepted, identified by their first bytes. Password protected files are refused.
- Scanned pages without a text layer have nothing to search.
- Whole word matching treats letters, digits and the underscore as parts of a word.
- Case insensitive search lowercases both the text and the query using your browser's rules.
Frequently asked questions
How do I search inside a PDF for a word?
Choose the PDF, wait while its text is read, type the word and press Search. The pages that contain it are listed with snippets.
Why does it find nothing in a scanned PDF?
A scan is a picture of text, not text. The tool reads only text that is stored in the file, so it needs a PDF with a text layer.
Does it search for phrases?
Yes. A phrase is matched as typed, and line breaks in the PDF count as spaces.
Is the PDF uploaded?
No. The file is read and searched in your browser and never leaves your device.