Searchable PDF

Make a scanned PDF or photo searchable

A scan or a phone photo of a page is a picture of text, not text. That is why search finds nothing in a scanned PDF even when the word is right there on the screen: the viewer is searching a photograph. This tool runs OCR (optical character recognition) on each page and adds the recognized words as an invisible, selectable text layer behind the page image. The page looks the same, and now it can be searched.

On the website, use Open to add JPG, PNG or HEIC images or an existing PDF. In the Isobj app, use Take Photo, Photos or Files. You can add up to 30 pages, then choose Create PDF.

Arrange and crop pages before recognition

Each page appears as a card. Drag cards to change the order, rotate a page a quarter turn at a time, or remove it. On the website you can also focus a page name and move it with the arrow keys; in the app, screen readers offer Move earlier and Move later on each page’s drag handle.

On the website, the crop button on any page opens it in a separate view where you select the document area. In the app, Take Photo uses the rear camera and opens the crop view right after the shot; pages added from Photos or Files don’t have a crop option. Text recognition and the finished PDF both use the selected area, so cropping away a desk edge or a thumb also keeps it out of the text. The tool does not correct perspective, so photograph the page flat and straight on.

Choose English or Spanish recognition

Set Document language to the language the page is written in before you create the PDF. It starts with Spanish when the site or app is in Spanish, and English otherwise. Clear, evenly lit printed text works best. Handwriting isn’t supported.

For compatibility, the hidden text stores an unaccented copy of each word and attaches the original spelling for viewers that use it. Depending on your PDF viewer, a search for canción may only match when you type cancion.

What happens to an existing PDF

On the website, each page of an imported PDF is rendered as an image. If a page already has a text layer, that text is kept and the page is not recognized again; pages without text, such as a scanned or image-only PDF, go through OCR. Cropping an imported page discards its existing text, and the cropped area is recognized instead. In the app, every imported page is rendered and recognized again by the phone’s text recognition, and you open one PDF at a time.

Because the pages become images, interactive form fields are flattened, links and bookmarks are not carried over, and any existing digital signature will no longer be valid. Every page in the finished PDF is sized so its longest side is 11 inches, which can differ slightly from the original paper size. A PDF longer than the 30-page limit imports its first pages that fit. Password-protected PDFs can’t be imported on the website.

Add a QR code or barcode label only if you need one

Page labels are off by default. Choose QR or Code 128 to place a label in one corner of every page. Each label encodes a document ID, the page number, the total page count and an optional case or file reference. Labels are added after recognition, so they don’t end up in the searchable text.

A random document ID is filled in for you, and you can change it to 1–64 letters, numbers, dots, underscores or hyphens. The reference can be up to 64 characters, and the label size can be set from 10% to 24% of the page. Anyone holding the page can scan the label, so keep names, account numbers and other personal details out of it.

Check the result, then save it

OCR can misread names, dates, numbers and faint printing, so compare anything important with the page image before you file or share the PDF. When pages were recognized, the website also shows an average OCR confidence. Both the website and the app list any page where no text was recognized; those pages can’t be searched, usually because the image is unclear or the document language is wrong.

You get the searchable PDF, a plain-text file of the recognized text and a small JSON manifest that records the pages and settings. The website downloads them; the app saves them to Files or shares them. The result is still a set of page images with hidden text, not an editable document. If you need to change the wording, PDF to Word turns a PDF into an editable Word file.

Your pages stay on your device

Recognition and PDF creation run locally. The website loads its OCR engine and English or Spanish language data from this site when recognition is first needed. The app uses the phone’s built-in text recognition: Apple Vision on iPhone and Google ML Kit on Android. Your pages and the recognized text are not uploaded. I built it this way because the documents worth making searchable, such as contracts, statements and forms, are usually the ones you would rather not hand to a conversion website.

Frequently asked questions

How does the PDF become searchable?

The tool runs OCR on each page and places the recognized words in an invisible, selectable text layer behind the page image. If no text is recognized on a page, it tells you that the page cannot be searched.

Which recognition languages are available?

English and Spanish. The selection starts with Spanish when the site or app is in Spanish and English otherwise; you can change it before creating the PDF.

Do I need to add a QR code or barcode?

No. Page labels are optional. If selected, each label contains a document ID, page number, total page count and an optional case or file reference.

What happens when I import a PDF?

Its pages are rendered as images on your device. On the website, existing text is kept and only pages without text are recognized; the app recognizes every page again. Interactive form fields are flattened, and existing digital signatures will no longer be valid.

Can I edit the text after OCR?

No. The PDF keeps each page as an image with hidden text that you can search and select, and you can also download the recognized text as a plain-text file. For an editable document, use PDF to Word.

Are my files or recognized text uploaded?

No. On the website, recognition and PDF creation run in your browser with an OCR engine and language data served by this site. The app uses the phone's built-in text recognition. Your pages and the recognized text stay on your device.

What are the current limits?

Up to 30 pages from PDFs, JPGs, PNGs or HEIC images, with 25 MB per source file and 100 MB total. Larger PDFs import the first pages that fit. Password-protected PDFs can't be imported on the website. Handwriting and perspective correction are not supported.