Search for words in a PDF, and fix it when nothing is found

How to search a PDF on any device

On a computer, press Ctrl+F (Cmd+F on a Mac), type the word and press Enter to go to the next match. This works in Chrome, Edge, Firefox and Safari when they show a PDF, and in Preview and Adobe Acrobat Reader too. On a phone, look for a magnifier icon, or a Find option in the app's menu.

If the app on your phone has no search, or you can't find it, open the PDF in our viewer. It runs in the browser, so there is nothing to install, and the file is never uploaded.

Searching with our PDF viewer

1. Open the PDF Viewer, tap Open a PDF and pick your file.
2. Press the magnifier in the toolbar. On a computer, Ctrl+F (Cmd+F on a Mac) does the same while the viewer is on screen.
3. Type a word or a phrase. The search starts a moment after you stop typing, or at once when you press Enter.
4. Read the count, like 3 of 42: the match you are on, and how many there are. Every match is marked in yellow, and the current one in orange.
5. Press Enter or the down arrow for the next match, and Shift+Enter or the up arrow for the one before. After the last match it goes round to the first. Press Esc to close the search.

What it matches

• Capitals don't matter. Rent, rent and RENT are all found.
• Accents don't matter on letters like é, so cafe finds café. Hindi and other Indian scripts keep their vowel signs, because in these scripts a sign changes the letter itself.
• A phrase is found even when it breaks across two lines, and extra spaces count as one.
• It finds the letters anywhere, even inside longer words. A search for rent also finds rental, current and parent. To narrow it down, add the next word: monthly rent. Spaces you type at the start or end are left out, so they won't help.
• It starts at the page you are reading and goes on through the whole file, so the first match you see is on the page you are on or after it.

How fast it is

In our tests in Chromium, a search for "trace" in a 14-page research paper found all 416 matches in about half a second. A word in a 117-page maths textbook took 2 to 3.5 seconds. A second search in the same file took about a tenth of a second, because the text of each page is read only once. We checked the counts against another PDF program, and they matched. Phones are usually slower than our test computer. The full details are in the viewer's guide to search.

Tips for finding the right place fast

• Search the rarest thing you know. In an exam result or merit list, search your roll number, not your name: names repeat and can be spelled in different ways. In a policy, search the policy number. In a bank statement, search the cheque number, the UPI ID or the shop's name.
• Type numbers the way the PDF prints them. 25,000 and 25000 are different to a search. So are ₹, Rs. and INR. If one form finds nothing, try the other.
• Try dates in more than one form. The same day can be printed as 05/03/2026, 05-03-2026, 05-Mar-2026 or 5 March 2026. Look at how the first page writes dates, and search the same way, or search only the part that stays the same, like Mar-2026.
• Search part of a long word. A search for reimburs finds reimburse, reimbursed and reimbursement in one go.
• Leave out apostrophes and quotes. Many PDFs use a curly apostrophe, which is a different character from the straight one on your keyboard. Search tenant instead of tenant's.
• Use headings in long documents. In an insurance policy, search Exclusions or Waiting period to land on the right section. If the file has bookmarks, the side panel takes you to a chapter faster than search.
• Several PDFs to search? Join them into one file with Merge PDF, then search the merged file in one go.

Why PDF search finds nothing

When search finds nothing for a word you can plainly see, the PDF almost always stores something different from what is drawn on the page. Search reads the stored text, not the picture. These are the usual causes, the most common first.

The page is a scan or a photo

A scanned page is a picture of paper. You can read it, but there are no letters inside the file for search to find. Office scanners, phone scanning apps and photos turned into PDFs all make files like this, unless they also ran text recognition (OCR: software that reads the letters in a picture and stores them as text). Our viewer says No text when it finds none at all. In a long scan, it stops after 20 pages without text and asks before searching the rest, because looking through every picture is slow.

The text was turned into a picture or shapes

Some PDFs only look typed. A document "printed" to PDF as an image, a form that was flattened, a page made from a screenshot, or a design file whose letters were turned into outlines for the printer: all show text but store none.

The font stores letters in its own code

Some PDFs keep their letters in a code that doesn't match what you see. Documents typed with older Hindi fonts, such as Kruti Dev, are the best-known example. The page shows correctly, but the stored text is a jumble of other letters, so a search for a Hindi word finds nothing, and copied text comes out as strange symbols. No viewer can search such a file properly.

The word is split at the end of a line

When a word breaks with a hyphen, like pay- at the end of one line and ment on the next, the file stores two pieces. A search for payment won't find it. Search for pay or ment instead. In our viewer, a phrase that simply wraps onto the next line without a hyphen is found as normal.

The phrase crosses columns, table cells or pages

A PDF stores text in its own order, which is not always the order you read it. In two columns, or in a table read row by row, a phrase you see together may be stored apart. Our viewer also searches one page at a time, so a sentence that runs from the bottom of one page to the top of the next isn't found as one phrase. Search a shorter piece of it.

The letters are stored with spaces between them

Headings with wide letter spacing are sometimes stored as single letters with gaps, like T O T A L. A search for total then finds nothing on that heading, though it still finds the word in the normal text.

A character looks the same but isn't

A curly apostrophe and a straight one, the rupee sign and Rs., a short dash and a longer one: each pair looks alike but is stored differently. Our viewer already treats joined letters, like the fi some fonts print as one sign, as their plain letters, but it can't guess that ₹ means Rs.

The text is in a form box or a comment

Text typed into the boxes of a form, and notes added as comments, are usually stored apart from the text of the page. Many viewers, ours included, don't search them.

You searched the web page, not the PDF

On our viewer's page, Ctrl+F opens the PDF search only while the viewer is on screen. Further down the page it is your browser's own search again, which looks at the page's words, not the PDF's. Scroll up to the viewer, click on a page and press Ctrl+F again.

How to tell if a PDF has real text

Before you try every search trick, spend a minute finding out what kind of PDF you have. Any of these tests will tell you.

1. Try to select a line. On a computer, drag across a line with the mouse. On a phone, press and hold a word. If the words light up one by one, the PDF has text. If nothing happens, or the whole page gets selected as one block, the page is a picture.
2. Search a word you can see on the first page. If it isn't found, and our viewer shows No text, there is no text at all.
3. Copy a sentence and paste it into a note. If it pastes as the same words, the text is good. If it pastes as boxes or strange letters, the font stores its letters in a code of its own.
4. Zoom in a long way, best on a computer. Real text keeps clean edges. A scan turns blurry and speckled.
5. Look closely. A page that is slightly tilted, grey, shadowed at the edges, or has a stamp or handwriting on it is usually a scan.
6. Compare the size with the page count. A text PDF is small for its pages, and a scan is much bigger. In our test files, a 14-page research paper was 992 KB, while a 42-page colour scan was 47.8 MB.

Some PDFs are a mix. A job application can have typed pages followed by scanned certificates and an Aadhaar copy. Search finds words on the typed pages only. If the first 20 pages of a long file have no text, our viewer pauses: press Search every page to go on.

Some scans carry hidden text. If OCR was run on a scan, the text it read sits invisibly under the picture, and search uses it. Recognition makes mistakes, though: a 0 read as an O, an l as a 1, rn as m. If a word on a scan isn't found, search a shorter part of it.

What to do when the PDF has no text

To be clear: our tools don't do text recognition. Our viewer shows a scan sharply and lets you zoom and turn it, but it can't make it searchable. Here is what does work.

Get a copy that has text

This is the best fix, because nothing has to be guessed. Ask the sender for the PDF saved straight from the program it was written in, not a scan of a printout. Download your bank statement from net banking instead of scanning the paper one. Save a form or a receipt from the website as a PDF rather than taking a photo of the screen. These copies have real text, and every search works.

Use text recognition elsewhere

• Google Drive can open an uploaded PDF or image with Google Docs and pull out its text.
• On a phone, Google Lens on Android and Live Text on an iPhone can pick out the text in a photo or a screenshot of a page. That is handy for a page or two.
• Desktop PDF software with OCR, such as the paid version of Adobe Acrobat, can make a whole scanned PDF searchable.
• Next time you scan, look in your scanning app's settings for an option called OCR or text recognition. Many apps can store the text as they scan.

Two warnings. First, most of these send your file to a server or keep it in an online account, which matters for an Aadhaar card, a bank statement or a medical report. Second, recognition makes mistakes, so check names and numbers against the page picture before you rely on them.

Find your way without search

A scan can still be read quickly. Open the side panel in our viewer to see small pictures of every page and tap the one you want. Fit page shows a whole page at once, so you can flick through looking for a table or a stamp. Type a page number in the page box to jump straight there. And if you need to know what changed between two scanned versions, Compare PDF can compare how the pages look.

Copying the text you found

Once search has taken you to the right place, you can copy the words from any PDF that has text. In our viewer the text of each page lies invisibly over the picture of the page, so you select it as you would on a web page. On a computer, drag across the words and press Ctrl+C (Cmd+C on a Mac). On a phone, press and hold a word, drag the handles over what you need, and choose Copy.

Copied text keeps the words, not the look. Tables come out as plain lines of text, so check the figures after you paste them. Make a note of the page number too: the page box shows it, and it makes the place easy to find again. Our viewer only reads, so it can't save highlights or notes into the file.

Questions people ask next

Can I search for whole words only?

Our viewer finds the letters anywhere, even inside longer words. To narrow the search, type two words together, like notice period.

Is PDF search case-sensitive?

Not in our viewer: capital letters don't matter. Most other PDF readers also ignore them unless you turn on an option.

Can I search a Hindi or Tamil PDF?

Yes, if the PDF stores its text in the standard way (Unicode). Type the word with its vowel signs as printed. PDFs made with older Hindi typing fonts can't be searched properly by any viewer.

Can I search several PDFs at once?

Our viewer opens one PDF at a time. Join them first with Merge PDF, then search the joined file.

Can I search a password-protected PDF?

Yes. Type the password when the viewer asks for it, and search works as normal. The password is used only on your device.

Are my searches sent anywhere?

No. The search runs in your browser, and the viewer keeps no record of your searches or your file.