Convert PDF to Non Searchable PDF: Flattening Explained

To convert a PDF to a non searchable PDF, turn every page into a picture and build a new PDF from the pictures. The step is called flattening to images, or rasterizing. Afterwards no word can be selected, copied or found. Work on a copy, test the result, and expect the recipient to lose search.

Why would anyone want a non-searchable PDF?

The sound reason is certainty about content: once each page is a picture, the file holds what the page shows and nothing behind it. People also flatten to freeze a document's appearance, to drop a misleading text layer, or to get past an upload system that rejects forms and scripts.

Reasons that hold up:

• After redaction. A page made into an image cannot carry words under a black box. The post on hidden text in a PDF explains why a drawn box alone does not remove them.
• To drop a bad text layer. A scan with a poor layer from an old copier returns wrong search results. Stripping the layer and running OCR again can be cleaner than stacking a new layer on the old one.
• To burn in markings. Stamps, annotations and filled form fields become part of the page and can no longer be moved or edited.
• To settle a display problem. A page with unusual fonts or layers that prints differently on different machines looks the same everywhere as an image.

Reasons that do not:

• To stop copying. Anyone can run OCR on your image-only file and have text again. Flattening inconveniences an honest reader and delays nobody else.
• As security. A flattened file is as readable by eye as it ever was. It is not protected in any sense.

What flattening a PDF to images removes and what it keeps

Flattening to images removes every character of text, visible or hidden, along with fonts, links, bookmarks, form fields and comments as live objects. It keeps the look of each page at the resolution you choose. Anything that was visible, including a redaction box or a sticky note, becomes part of the picture.

Two side effects follow. The file usually grows, because pictures take more space than characters. And sharp vector text becomes a grid of dots, so small print can soften. Pick a resolution high enough to read footnotes; the post on the best DPI for OCR explains what the number means.

Take a complaint through this conversion and you can no longer search the plaintiff, the cause of action, the damages demand, or a citation in the margin. The picture of the complaint stays on the page. The person who receives it reads by eye, or runs recognition again, before those words can be found.

Be careful with the word "flatten". It has a second, weaker meaning. The federal courts' guide to PDF functionality for public filers in NextGen CM/ECF tells filers to flatten a rejected PDF by printing it to a PDF printer. That kind of flattening merges forms and other interactive parts into the page. It normally leaves ordinary text as text, so the result is still searchable. A PDF editor's "flatten" command often does the same. Only a method that produces images gives you a non-searchable file.

Document properties are a separate matter. A new file gets new properties from the program that made it, and those may still include a title or author name. Check them.

Ways to convert a PDF to non-searchable pages

There are four practical routes: export the pages as images and recombine them, print with an "as image" option to a PDF printer, print on paper and scan, or use an online converter. The first two keep the most quality. The last sends your file to someone else's server.

Export and recombine. This works with most PDF programs:

1. Save a copy of the PDF under a new name, such as Exhibit-C_image-only.pdf.
2. Export every page as an image. Choose PNG or TIFF at 300 dpi for text pages.
3. Check the image count equals the page count.
4. Combine the images, in order, into a new PDF.
5. Make sure no "recognize text" or "make searchable" option is switched on while you combine. Some programs run OCR by default and would put the text straight back.

Print as image. Many print dialogs have an advanced setting called "Print as image". Select it, choose a PDF printer as the destination, and print. Without that setting, printing to PDF is the weaker kind of flattening described in the section on what flattening removes.

Print and scan. Paper has no text layer. Print the pages, scan them with the scanner's searchable-PDF option turned off, and you have an image-only file. Quality drops a little, and this is the slowest route, but nobody can doubt the result.

Online converters. Searches for "convert pdf to non searchable pdf online" lead to free websites that do the export-and-recombine step for you. Before you use one on a client document, read what the site keeps and for how long. The questions to ask are in the post on whether it is safe to upload client documents to an OCR tool. A file you are flattening because of what it contains is a poor candidate for an unknown server.

None of Schema Law's eight export formats is an image-only PDF. It works in the other direction, reading pages into text and exporting a searchable PDF with a hidden text layer; the export formats on the homepage show what it produces.

How do you confirm no text is left in a flattened PDF?

Try to get text out of the file by every ordinary means. If selecting, searching, copying and exporting to text all return nothing on every page, the file is image-only. One page of live text among two hundred images is enough to fail, so test the whole file.

One trap can fool the first test. Some PDF viewers and operating systems recognize text in pictures on the fly, so you can select words in an image that contains no text layer at all. That text is being produced by your viewer at that moment and is not stored in the file. The export-to-text test, in a program without that feature, settles it.

Then open the document properties and look at the page count and the font list. A flattened file that still lists embedded fonts did not flatten the way you intended.

What a non-searchable PDF costs the person who receives it

The recipient loses search, copy and paste, working links and bookmarks, and screen-reader access. They get a larger file. If they need the text, they will run OCR themselves and work from a version with recognition errors that you could have avoided.

Count those costs before you flatten a whole document to solve a problem on one page:

• Search. A 300-page production with no text cannot be searched for a name. What a text layer gives the reader is set out in the post on the searchable PDF.
• Accessibility. Screen-reading software has nothing to read in a picture of a page.
• Filing rules. A court's rules may require text-searchable filings, and the details vary by court. Idaho's Rule 6 of the Rules for Electronic Filing and Service, for example, calls for text-searchable PDF and makes an exception for scanned attachments and exhibits. The section on whether courts require text-searchable PDFs collects more examples. Read your own court's rule before you file an image-only document.
• Agreed production formats. A discovery agreement or order may specify searchable text. An image-only production can breach it.

There is usually a middle path. Flatten to remove everything hidden, then run OCR on the flattened file. The new text layer is made from the picture, so it holds only what a reader can already see, redaction boxes included. You get a clean file and the recipient gets search. Run the leftover-text checks between the two steps, and again search the final file for every term you meant to remove.