# Making Scanned Documents Searchable and Editable: Two Files

- Author: Schema Law (https://www.schemalaw.com/)
- Published: 2026-03-22T09:58:08Z
- Updated: 2026-10-11T19:22:20Z
- Category: Searchable PDFs
- Canonical: https://www.schemalaw.com/blog/making-scanned-documents-searchable-and-editable

> Making scanned documents searchable and editable takes two files: a searchable PDF as the record and a Word copy to edit. How to make, use and pair them.

Making scanned documents searchable and editable takes two files, not one. A searchable PDF keeps the scanned pages and lets you find words on them. A Word file holds the same words as text you can change. Make both from one conversion, keep the PDF as the record, and do your editing in Word.

One file that does both sounds tidier. The idea breaks down as soon as someone changes a word, and the reason is worth a minute before you set up a matter folder around it.

Check the searchable PDF before you rely on either copy. Search the plaintiff, a damages figure, and the contract clause that states an obligation. Compare that line with the page image, then find the same words in the Word file. Where the editable line and the picture disagree, the picture is the line you checked.

## Can one file make a scanned document both searchable and editable?

No single file makes a scanned document both searchable and editable in the way legal work needs. A searchable file must keep the scanned page as it is. An editable file must let the words move when you change them. A picture cannot reflow, so the two jobs go to two files.

Can you make a scanned PDF searchable? Yes, and the result still looks exactly like the scan. Can you then edit that same file? Only by breaking the thing that made it trustworthy, as four attempts at a single file show.

| Attempt at one file | What you get | Where it fails |
| --- | --- | --- |
| Edit the hidden text of a searchable scan | The picture shows the old words; search finds the new ones | The file now says two things at once |
| Type over the scanned page in a PDF editor | New type patched onto a picture | The patch does not match the print, and the page is no longer the document that was scanned |
| Keep only the Word conversion | Fully editable text | The signatures, stamps, handwriting and page layout the paper showed are gone |
| Save the Word conversion back to PDF and discard the scan | A clean, searchable, typed PDF | It is a new document; nothing in it is a picture of the paper |

The first row is the subtle one, and the post on searchable PDFs sets out the distinction under [whether a text searchable PDF is the same as an editable one](https://www.schemalaw.com/blog/searchable-pdf#is-a-text-searchable-pdf-the-same-as-an-editable-one). The second row has its own guide: the post on [editing a scanned PDF](https://www.schemalaw.com/blog/editing-scanned-pdf) covers changing the scan itself and when that is defensible.

Two files remove the conflict. Nobody has to choose between a page that can be trusted and text that can be changed, because each file does one of those and makes no claim to the other.

## The searchable PDF as the record and the Word file as the working copy

The searchable PDF is the record: it shows what the paper showed, and its text exists so you can find things. The Word file is the working copy: it carries the recognized words in a form you can revise, and it carries no authority about what the paper said.

| | Searchable PDF | Word file |
| --- | --- | --- |
| What a page looks like | The scan, unchanged | A fresh layout built from recognized text |
| Signatures, stamps, handwriting | Visible, as pictures | Absent, or present as pasted pictures at best |
| Pages break where the paper's pages break | Yes | No, text reflows when fonts or margins differ |
| Recognition errors | Hidden behind a correct picture | Visible in the text you read |
| Safe to change | No | Yes, that is its purpose |
| Answers "what did the document say?" | Yes, by looking at the page | No, it answers "what did the software read?" |

Two rows deserve attention.

**Page numbers.** A citation to page 14 of the agreement means page 14 of the paper, and the searchable PDF shows that page as it was printed. The Word file may put the same clause on page 11. Never take a page reference from the working copy.

**Where the errors show.** A misread word in a searchable PDF sits behind a correct picture, so a reader sees the right word and only a search goes wrong. The same misreading in Word is on the page in front of you, and it will be copied into whatever you draft from it. That is the reason to proofread the Word file against the PDF before anyone relies on it, and the post on [checking OCR output against the original](https://www.schemalaw.com/blog/why-is-ocr-output-full-of-errors) describes how.

Mark the working copy so its status is plain to someone who opens it cold. The post on converting a signed agreement to Word gives wording for that under [labeling the Word file so nobody mistakes it for the executed agreement](https://www.schemalaw.com/blog/convert-signed-agreement-to-word#labeling-the-word-file-so-nobody-mistakes-it-for-the-executed-agreement).

## Which copy do you cite, send and file: the searchable one or the editable one?

Cite from the searchable PDF, send whichever copy fits what the recipient will do with it, and treat anything that goes to a court or an opponent as governed by their rules. As a working rule, tasks about what the document says use the PDF, and tasks that produce new text use the Word file.

| Task | Copy to use | Reason |
| --- | --- | --- |
| Find every mention of a name or term | Searchable PDF | Search lands on the page as scanned |
| Quote a clause in a brief or letter | Find it in the PDF, copy from either, proofread against the page picture | The picture is the authority for the wording |
| Give a page or paragraph reference | Searchable PDF | Its pages are the paper's pages |
| Highlight and annotate while reading | A copy of the searchable PDF | Marks sit on the real page |
| Draft an amendment or a redline | Word file | You need text that moves |
| Reuse a clause as a precedent | Word file, after proofreading | Clean text without the old document's look |
| Send to a client "for their records" | Searchable PDF | They want the document, with signatures |
| Send to co-counsel to edit | Word file, with the PDF beside it | They can check the text against the pages |
| File with a court or produce to the other side | Check the applicable rule or protocol first | Requirements differ by court and by case |

On the last row, do not assume. Courts set their own requirements for filed PDFs, and the post on [searchable PDFs for court filing](https://www.schemalaw.com/blog/searchable-pdf-for-court-filing) collects what to look for in your own court's rules. A production may be governed by an agreed protocol that names the format.

For reading with a highlighter, make a third file: a copy of the searchable PDF kept for markup, so the clean searchable copy stays clean. Whether a highlight will follow the words depends on how the text layer was built, and the post on [converting a PDF to a highlightable PDF](https://www.schemalaw.com/blog/convert-pdf-to-highlightable-pdf) explains what to test.

### A question to ask before sending the Word file

Ask what the recipient might do with it that you would not want. A Word file invites editing. If an edited version comes back, or travels onward, can anyone still tell what the signed text was? Sending the PDF alongside the Word file answers that. Sending Word alone leaves the recipient with nothing to check it against.

## Producing the searchable and editable versions from one conversion

Produce both versions from a single recognition pass, so the hidden text in the PDF and the body of the Word file contain the same reading of the page. One upload with two exports does that. Two separate conversions in two tools give you two different sets of mistakes.

The practical value is in the checking. When both files come from the same pass, a misread name you spot while proofreading the Word file tells you something about the PDF too: a search for the correct spelling will miss that spot. Note the misreading once and it serves both files.

In Schema Law the sequence is short, and the [export formats section of the homepage](https://www.schemalaw.com/#formats) puts it in one line: "One upload, as many exports as you need."

1. Copy the scan as received into a folder you will not work in.
2. Sign in and upload the PDF, PNG or JPEG.
3. When the pages have been read, look through them on screen against the original.
4. Export Searchable PDF. Save it with the record suffix you use.
5. Export Word from the same document. Save it with the working-copy suffix.
6. Proofread the Word file against the PDF, starting with names, dates and amounts.

Both formats are on every plan, as are the other six. The Free plan takes one document a day of up to three pages, which is enough to try the pairing on a real letter or a short order. Longer documents need a plan with a larger page allowance, and the [pricing page](https://www.schemalaw.com/pricing) gives the page limits for each. The homepage also says you can "export the same document again in a different format" later, so the Word copy can wait until someone needs it.

### If you use a different tool

The same principle holds in any tool, and the guide on [how to make a scanned PDF searchable](https://www.schemalaw.com/blog/how-to-make-a-scanned-pdf-searchable) has the basic procedure. If your desktop program recognizes text inside the PDF, save the searchable PDF first, then export Word from that saved file so the export reuses the stored text. Check afterwards that it did: pick one misreading in the Word file and search the PDF for the same wrong spelling. If the PDF has it too, the two files share a pass. If not, the program recognized the pages a second time and you have two readings to keep track of.

### What if you only need one of them today?

Make the searchable PDF regardless. It comes from the same upload, and it is the file you will want when a question comes up about the text of the Word copy. The Word file is the optional one.

## Keeping the searchable and editable versions of a scan paired in the matter file

Keep the pair together by giving both files one base name, adding a suffix for the role each plays, and storing them where a person who finds one will see the other. Record in the Word file itself which PDF it came from, so the link survives a move or a rename.

### Names that sort together

Suppose the fictional firm acting for Dunmore Freight scans a signed warehouse lease. A set of names that keeps the family in one place:

| File | Role |
| --- | --- |
| `2021-03-03 Lease Harlow-Dunmore SCAN.pdf` | The scan as it came off the copier |
| `2021-03-03 Lease Harlow-Dunmore SEARCHABLE.pdf` | The record you search and cite |
| `2021-03-03 Lease Harlow-Dunmore TEXT working copy.docx` | Recognized text, proofread, never edited further |
| `2021-03-03 Lease Harlow-Dunmore TEXT working copy, amend draft v1.docx` | A draft built from the working copy |

The third and fourth rows are separate on purpose. Keep one Word file that is nothing more than the proofread conversion, and start every draft from a copy of it. Otherwise the first amendment draft quietly becomes the only Word version of the lease, with changes nobody can tell apart from the original wording.

### A provenance line inside the Word file

Put one line in the header or on a cover page of the working copy:

> Text converted by OCR from "2021-03-03 Lease Harlow-Dunmore SEARCHABLE.pdf". Proofread against that file on 9 April by T. Okafor. Not the executed document.

File names change and folders get reorganized. A line inside the document goes wherever the document goes.

### In a document management system

If your system lets you relate documents to each other, relate the three. If it profiles documents by type, give the scan and the searchable PDF a type that says "executed" or "as received" and the Word file a type that says "working copy" or "draft". Avoid saving the Word file as a new version of the PDF, or the reverse. Versions imply one document changing over time, and these are different things.

### When the scan is replaced

If a better scan arrives, such as a cleaner copy from the client, redo the whole family from it: new searchable PDF, new Word conversion, new proofread. Retire the old set together and say so in the file note. A Word file proofread against a PDF that no longer exists in the folder cannot be checked by anyone.

## When edits to the Word copy must go back into a new PDF

When edits to the Word copy have to circulate as a PDF, save the Word file as a new PDF and treat it as a new document. It does not replace the scan or join the scanned record. It has typed pages and no signatures, and its name should say what it is.

Saving from Word gives you a PDF with real text, so it is searchable without OCR. The details, including the difference between saving and printing to PDF, are in the post on [converting Word to a searchable PDF](https://www.schemalaw.com/blog/convert-word-to-searchable-pdf).

### The round trip, step by step

1. Start the draft from a copy of the proofread working file.
2. Make the edits with tracked changes on, so the difference from the converted text is recorded.
3. Save the result as a PDF from Word.
4. Name the PDF for what it is: `Lease First Amendment DRAFT v1.pdf`, not anything that resembles the scan's name.
5. Search the new PDF for one word from the first page and one from the last.
6. File it with the drafts, not with the executed documents.

### Three mistakes the round trip invites

- **Swapping pages.** Do not replace a page of the scanned record with a retyped page from Word, even to fix an obvious typo in the original. The record shows the paper, typos included. The post on editing a scanned PDF discusses the narrow cases under [whether you should ever alter the scan of a signed or filed document](https://www.schemalaw.com/blog/editing-scanned-pdf#should-you-ever-alter-the-scan-of-a-signed-or-filed-document).
- **Calling the typed PDF a copy of the agreement.** A PDF made from the Word conversion is a transcription. If it is sent as "a copy of the lease", the recipient may take recognized text for signed text.
- **Losing the loop.** When the amendment is signed on paper and scanned, it comes back as a picture. It needs its own searchable PDF and, if anyone will draft from it, its own working copy. The family of files for the amendment is separate from the family for the lease.

### How it looks over the life of one document

In the Dunmore example, the lease is scanned in March and its searchable PDF and working copy are made that week. In June the client wants to extend the term. The associate copies the working file, drafts the first amendment with tracked changes, and saves a PDF of the draft for the landlord's counsel. After two rounds the amendment is signed, scanned and converted as a document in its own right. The folder now holds two records, two working copies and a trail of drafts, and each file's name says which of those it is. Nothing in that sequence required anyone to edit a scan.

## Try it on a real document

Upload a scanned PDF and export it as Word, a searchable PDF or plain text on the free plan. Your documents are never used to train an AI model. [Start free](https://www.schemalaw.com/api/auth/google)