Scan to Searchable PDF: Paper to a Findable File in One Pass

To scan to searchable PDF in one pass, choose the searchable or OCR file type on the scanner or in its software before you press start, feed a prepared stack, and send the result to a folder you control. Then open the first file and search it before you scan anything else.

One pass saves a step only when the file that comes out can be trusted. A device that recognizes text while it scans gives you no chance to look at the picture first, so the care moves to the two ends of the job: what you do to the paper before it goes in, and what you test when the first file comes out.

On the first file the copier writes, search the party names and a docket number in the caption. If the sheet is a hearing notice, search for a case management conference and for the clerk of court. Compare each hit with the page image before the rest of the stack goes through. A notice that fails this test means you change the device settings and scan that item again.

Can an office scanner turn paper into a searchable PDF in one pass?

Yes, an office scanner can make a searchable PDF in one pass if it or its software has text recognition switched on. The recognition runs in one of three places: inside the machine, in the scanning program on the attached computer, or on a server the machine sends its images to. Find out which.

Where the recognition runs decides what settings you get, how long a job takes and where the page images travel. Here is how to tell which arrangement you have.

The third arrangement deserves a question of its own when the paper is client material. If a scan leaves the building to be recognized, the firm should know which service receives it and on what terms. Your IT contact or the device administrator can answer from the configuration.

What the panel's searchable PDF setting does inside the machine, step by step, is explained in the post on what the searchable PDF button on an office copier does. The short version: the device scans each sheet, runs its own recognition, and hides the text behind the pictures with nobody reviewing it.

If the option is missing

A missing option has three possible causes: the model does not have the feature, the feature is an unlicensed add-on, or an administrator turned it off. Ask before you assume. If there is no built-in recognition, scan to an ordinary PDF and add the text afterwards. The guide to creating a searchable PDF by the route that fits what you have sorts those routes.

Preparing a paper document for a scan-to-searchable-PDF run

Prepare the paper so that the stack feeds once, in order, without stopping. Remove every fastener, flatten folds, repair tears, put the pages in order facing the same way, and mark where one document ends and the next begins. A one-pass scan has no second chance short of scanning again.

Work through the stack in this order:

1. Count the sheets and write the number down. You will compare it with the page count of the file.
2. Remove staples, clips and binder tabs. A staple that reaches the feeder can scratch the glass strip, and that scratch then draws a line on every later page.
3. Unfold and flatten. Smooth dog-eared corners. A folded corner hides text and can drag a second sheet through.
4. Deal with small and fragile items. Receipts, envelopes, torn pages and anything with sticky notes go on the glass by hand, or are photocopied onto a full sheet that can be fed. Move a sticky note off the text it covers and scan the page twice if both the note and the text matter.
5. Check the order and orientation. Every page face up or face down as the feeder requires, top edge leading the same way.
6. Mark the document breaks. Decide whether each document becomes its own file or the stack becomes one.

Separator sheets

A separator sheet is a page placed between documents so that software, or a person, can find the boundaries afterwards. Some scanning programs can split a stack into files at a blank page or at a printed barcode sheet. Check whether yours can before you rely on it. If it cannot, a brightly colored sheet with the next document's name written large still helps, because you can see it in the thumbnail view and split the file there by hand.

Two cautions. A separator sheet becomes a page in the file unless something removes it, so your page count will be the sheet count plus the separators. And a setting that drops blank pages may remove blank separators before they can be used as split points. Decide which behavior you want and test it on a short stack.

Settings to fix before the first sheet goes in

Resolution, color mode and file naming are chosen at the panel or in the scan profile. They are set out in the guide to scanner settings that make OCR work. Set them once, save them as a named profile or preset if the device allows it, and use the same preset for the whole job. If the panel has a language list for recognition, set that too.

Scan to email, scan to folder, scan to computer: where the searchable file lands

Scan to email sends the file as an attachment, scan to folder writes it to a network or cloud location, and scan to computer hands it to software on a workstation. For client paper, a folder the firm controls is the safer destination, because the file arrives whole and stays where the matter is kept.

Each destination has a failure of its own.

Email

Email is convenient for a five-page letter. For a long or color document, find out the attachment limit on the device and on your mail system before the job, because a device may be set to split a large scan into several messages or to fail the send. If the parts arrive as separate files, the document is no longer one searchable PDF. Email also leaves a copy of a client document in the sent items of the device account and in the inbox of whoever received it.

Folder

Scan to folder is the better default for matter documents. Ask two things about the folder: who can open it, and how files are named. A shared "Scans" folder that every employee and every temporary worker can read is the wrong place for a client's medical records to sit overnight. Move each file to its matter folder the same day, and empty the landing folder.

Computer

Scan to computer puts the recognition, the naming and the destination under one program's profile. That is where a searchable option can live when the device itself has none. The post on scanner software that recognizes text at scan time looks at one such program from its maker's documentation.

Whatever the destination, rename the file when it lands. A device name such as SCAN_0042.pdf tells nobody what it is. If scanning the day's post becomes a regular task, the post on scanning incoming mail to searchable PDF as a daily routine lays out who does what.

Why does the copier's searchable PDF look fine and search badly?

The page picture and the hidden text are made separately, so one can be good while the other is poor. Two settings do the damage on a one-pass scan: compression that keeps the file small at the cost of letter shapes, and quick built-in recognition run with defaults that do not match the document.

Compression chosen for small files

A device set to produce compact files may use a high-compression PDF mode. Such modes store the dark text and the page background separately and squeeze each hard. At normal zoom the page looks clean, even cleaner than the paper. Zoomed in, small letters can show filled-in loops or broken strokes, and those are the shapes the recognition had to work from. The methods involved are explained in the post on lossy versus lossless compression.

Look on the panel for file type or compression choices with names such as compact, high compression or small size, and scan one page with that setting off to compare.

Recognition set up once and never revisited

The recognition inside a device runs with whatever was configured at installation. Check these in the panel or the administrator's page:

• Language. A recognition language that does not match the document produces wrong characters with no warning.
• Resolution. A low default suits email attachments and starves recognition of detail.
• Color mode. Pure black and white can drop faint print before recognition sees it.
• Automatic rotation. If it is off, sideways pages are stored with sideways text, or with none.

How to confirm the cause

1. Pick one page with small print, such as a page of contract boilerplate.
2. Scan it with the usual one-pass setting.
3. Scan it again as a plain PDF at a higher resolution with compact compression off.
4. In the first file, select the page's text and paste it into a blank document. Read it against the paper.
5. Run recognition on the second file with a separate tool and do the same.

If the second result reads better, the paper is fine and the device's defaults are the problem. Either change the preset or scan plain and add the text afterwards. The post on why OCR output is full of errors separates scan faults from software faults more fully.

Checking the first scan-to-searchable-PDF file before you feed the whole stack

Scan the first document, or the first ten sheets, and stop. Open that file at your desk and check the page count, the picture and the search before you feed anything else. A wrong setting found on ten pages means rescanning ten. Found after six hundred, it means rescanning six hundred.

Use this checklist on the first file:

• The page count matches the sheets fed, allowing for backs and separators.
• Every page is upright and nothing is cut off at an edge.
• No vertical line or streak runs down the pages.
• A search for an ordinary word on the last page finds it.
• A search for a surname and for a dollar amount or date finds each one.
• Text selected on one page and pasted into a blank document reads as sentences.
• At 300% zoom, small print is still made of clean letter shapes.
• The file landed in the right place under a name you can change.

Include the worst page in the trial on purpose. If the stack contains a faxed page, a carbon copy or a page of fine print, put one of them among the first ten sheets, so the test tells you about the hard pages as well as the easy ones.

A worked example

A paralegal at a fictional firm, Calloway Reyes LLP, has a 412-sheet client file in the matter of Thornbury Marine to scan before a deadline. She feeds the first twelve sheets as a searchable PDF to the matter's intake folder. The file has twelve pages, though three of the sheets were printed on both sides: the device was set to one-sided. A search for "Thornbury" finds four pages out of the nine that show the name. She turns on two-sided scanning, raises the resolution on the preset, rescans the twelve sheets and gets fifteen pages and nine hits. Only then does the rest of the stack go in.

When the whole job is done, repeat the page count for the full file and search once more on a late page. Checking a finished set file by file is the subject of the post on how to check whether a PDF is searchable. Keep the paper until those checks are done.

Two-sided originals, blank backs and mixed paper sizes in one pass

A one-pass scan goes wrong at the feeder in three ways: backs that are never scanned, blank backs that are kept or dropped by a setting nobody chose, and sheets of different sizes that are cropped or skewed. Each one changes the page count, so counting is how you catch them.

Two-sided originals

If any sheet in the stack has print on the back, scan the whole stack two-sided. A one-sided pass over a mixed stack silently loses every back page, and the searchable file then appears complete while missing half a contract's terms. How two-sided scanning works and how to tell from a file whether backs were captured is covered in the post on what duplex scanning is.

Blank backs

Scanning everything two-sided produces a blank page for every sheet printed on one side. A blank-page removal setting tidies that up, and it decides what counts as blank by how much ink it sees. A back that carries only a received stamp, a handwritten initial or a faint notation can fall under its threshold and vanish. For a file where every mark matters, leave removal off, accept the blanks, and delete true blanks by eye afterwards on a copy.

The arithmetic for your count:

Mixed paper sizes

A stack that mixes letter and legal pages, or full sheets and half-sheet receipts, needs a mixed-size or automatic size setting if the device has one. Without it, the scanner uses one size for everything, and the bottom of each legal page is cut off, taking signature blocks and notary certificates with it. The post on scanning legal size documents deals with cropped edges.

Small and odd-sized sheets also skew in the feeder, because the guides cannot hold them. Scan them on the glass and insert the pages into the file afterwards, or feed them in a batch of their own with the guides set to their width. The differences between feeder and glass are laid out in the post on flatbed versus sheet-fed scanners.

When one pass is the wrong plan

Some jobs are too large or too fragile for an office feeder: dozens of boxes, bound volumes, brittle paper. For those, the post on ordering searchable PDFs from a scanning service covers sending the boxes to a vendor instead. And where the device's own recognition keeps failing your first-file test, scan to a plain PDF and convert it separately. Schema Law takes a PDF, PNG or JPEG of up to 20 MB and exports a searchable PDF among other formats; the how it works section shows the steps.