Free OCR Scanner Apps: Does a Phone Scan Contain Text?

A free OCR scanner app sometimes gives you recognized text and sometimes gives you only a picture of the page. The scan can come out one of three ways: text written into the PDF, text kept inside the app, or no text at all. The file itself tells you which, and checking takes a minute.

The word "OCR" on a store listing does not settle it. The question for a matter file is narrower: when the PDF leaves the phone and lands in a folder at the office, can someone search it?

Does a free OCR scanner app put text in the file?

Sometimes. An app that advertises OCR may write the recognized words into the PDF it exports, may recognize them only for use inside the app, or may leave recognition out of the free version. Only the first outcome gives you a file that is searchable anywhere.

The second and third rows produce identical files. The difference shows only on the phone, where the app's own search box can find a word in the second case. That is why a scan can seem searchable to the person who made it and prove unsearchable to the person who receives it.

A PDF with pictures and a hidden layer of characters is called a searchable PDF. The post that defines what a searchable PDF is explains the format and its variants.

Text you can search in the app versus text saved in the PDF

Recognition can live in two places: in the app's own index, or in the file. Text in an index helps you find a scan inside that app. Text in the file travels with the PDF to every reader, folder and search tool that opens it afterwards.

Think of the index as a card catalogue the app keeps about your scans. The cards say which words appear on which pages. The scans themselves are unchanged pictures. Export one, and the cards stay behind.

Three signs that the text lives only in the app:

• Search works inside the app, and the same word is not found once the PDF is opened in a different program.
• The app offers "copy text" or "export as text" as a separate action, and the PDF export is described as an image or a scan.
• The app's help pages describe searching your library, and say nothing about the contents of the exported file.

The same split exists outside phones. A storage service can index a scan for its own search box without changing the file, a case examined in the post on Dropbox OCR. And a tool that copies recognized words to the clipboard gives you text for one paste and leaves the file as it was.

For legal work the file is what counts. A scan goes into a document management system, gets attached to an email, or is handed to co-counsel. None of those recipients has your app.

Search the exported PDF of a client's pleading before you call the job done. Keep the photograph of each page beside it, because the hidden text is only a reading. If that file may serve as documentary evidence, open the picture for authentication. A folder note of who exported it records the chain of custody.

Which free phone scanners say they save a searchable PDF?

Google says so for two of its Android tools. Its help pages state that scans made with the Google Drive app and with Files by Google are saved as searchable PDFs. Apple's scanning guide for iPhone does not address recognized text at all, so an iPhone scan is something to test.

Two cautions about the Google rows. The Drive page lets you save a scan as .pdf or .jpg, and a JPG is an image format with no place for a text layer, so choose PDF. And neither page says how good the recognized text is, what languages it covers or how handwriting is treated. "Searchable" is the vendor's word for the format, and it is not a promise about any one page.

For Apple, silence in that guide is not a statement that the text is absent. It means Apple's scanning guide makes no claim, and you find out by exporting a scan and checking it. The guide to scanning a document on iPhone covers getting the PDF out of Notes so you can.

For any other app, read the vendor's own help page for the words "searchable PDF" or "text layer", and note whether the feature belongs to a paid tier. The post on Genius Scan OCR works through one vendor's pages as an example, and the guide to choosing a free scanning app puts text in the file among its comparison criteria.

How do you check a phone scan for a text layer in one minute?

Send the PDF out of the app, open it in a different program, and search for a word you can see on the page. A hit means the file carries text. No hit on a page full of that word means the file is pictures only.

1. Export the scan as a PDF. Share it to email or save it to a folder. Do not test inside the scanning app, because the app's own search may be reading its index.
2. Open the file somewhere else. A computer's PDF reader is best. A different PDF app on the phone will do.
3. Search for a plain word from the middle of a page. Pick one that is printed clearly.
4. Try to select one line. Drag across it. Text highlights word by word. A picture does not.
5. Paste the line into a note. Read it against the page. Scrambled characters mean a layer exists and was recognized badly.

One page proves little about a forty-page scan. For a file you will rely on, run the fuller set of searches described under how to confirm a PDF is really searchable, which covers names, numbers and late pages.

When a free OCR scanner's text is not enough

A phone scanner's text layer serves for finding a word in a short document. It falls short when you need a table as rows and columns, an editable document, text from a long file you must be able to vouch for, or any format other than a PDF.

• Tables. A hidden text layer records words for searching. It does not give you a ledger you can sort or total. That takes an export built for tables, as the post on scanned PDF to Excel explains.
• Editing. A searchable PDF is still a picture of a page. To revise the wording you need the text as a document.
• Long files. Photographing sixty pages by hand produces sixty separate captures of varying quality, and a phone gives you no easy way to compare each page of text with its image.
• Other formats. The phone tools described here save a PDF or an image. Plain text for a brief, or structured output for a spreadsheet, needs a conversion step.
• No text at all. If the check found nothing, the file needs recognition from scratch.

In each case the phone scan is still useful as the page image. Keep it, and run recognition on it separately. Schema Law takes a PDF, PNG or JPEG of up to 20 MB, reads each page into text and tables, lets you view the parsed pages before you export, and offers the formats listed in the export formats section of the homepage. The free plan handles one document a day of up to three pages, which is enough to compare its output with your scanner app's on the same pages.

Whichever route you take, recognized text is a reading of the page and not the page. Keep the scan as the record, check names, dates and amounts against it, and say in the file which copy is which.