How to make a scanned PDF searchable
Open Make searchable, pick the language the scan is in, and press Make it searchable. The words are laid invisibly over each page, which looks exactly the same.
Make searchableFree. Nothing is uploaded.
How to do it
- Open Make searchable and add the scanned PDF.
- Pick the main language of the scan, and up to two more if it mixes languages.
- Press Make it searchable.
- Save the result and search it with your PDF reader’s find.
Languages
English is ready straight away. Twenty more can be picked, among them Malayalam, Hindi, Tamil, Telugu, Kannada, Bengali, Marathi, Gujarati, Punjabi and Urdu, as well as Arabic, French, Chinese and Japanese. Each is fetched from this site the first time you pick it, and the size is shown before you start. Pick only the languages actually on the page: each extra one makes the reading slower.
Indian scripts, honestly
The invisible layer inside the PDF can hold Latin letters only for now: the ones English, French, German, Spanish, Portuguese and Italian use. When you pick Malayalam, Hindi or another language written in a different script, everything read also comes out as a .txt file beside the PDF, so you can copy and search it there. The result says which file holds what.
Better scans read better
Straight, sharp, high-contrast pages read best. Scan in Black and white gives the reader its easiest job. Pages that already have text are left alone. To get the words out as plain text afterwards, use Extract text; for an editable document, PDF to Word.
A worked example
You have a 30-page scanned property file in English and Malayalam and need to find every mention of a survey number. Open Make searchable, add the file, pick English and Malayalam, and press Make it searchable. Save both results. The searchable PDF finds the English words and the numbers with your reader’s find, and the .txt file holds every word that was read, Malayalam included, to search in any text app.
Checking the result
Open the searchable PDF and use find for a word you can see on a page. If it is found and highlighted in the right place, the hidden layer lines up with the picture. Then select a sentence and paste it into a note: what pastes is what was read, which shows you how accurate the reading was for that page before you rely on it.
For names, account numbers and amounts, do not trust the reading blindly: a 1 can be read as an l, or a 0 as an O. Search for one or two numbers you know are on the page to see whether they come out right.
Common mistakes
Picking every language that might appear: each extra one slows the reading and makes mistakes more likely, so pick only what is really on the pages. Running it on a crooked phone photo, when straightening it with Scan first would read far better. And expecting handwriting to come out well; printed text is what the reader is made for.
Questions people ask
Is my scan uploaded to be read?
No. The reading happens in your browser. Only the language model comes from this site, once; your document goes nowhere.
Will it change how my pages look?
No. The words go in as an invisible layer; the pictures are not touched.
How accurate is it?
Very good on clean printed text, weaker on handwriting, faint copies and unusual fonts. Check anything important.
Can I pick more than one language?
Yes, up to three. More than that makes the reading worse, not better.
Does it work on a phone?
Yes, though reading every page is real work for a phone, so a long scan takes a while. A computer is quicker for big files.
Can I make a searchable PDF from photos?
Yes. Turn the photos into a PDF with Scan first, then make that searchable.
Your file stays on your device. Foolscap does the work inside your browser, so nothing is uploaded and there is nothing for anyone to keep. How you can check