# Convert PDF to Text Searchable Files a Folder Search Finds

- Author: Schema Law (https://www.schemalaw.com/)
- Published: 2025-09-10T17:00:43Z
- Updated: 2026-10-11T19:22:20Z
- Category: Searchable PDFs
- Canonical: https://www.schemalaw.com/blog/convert-pdf-to-text-searchable

> Convert PDF files to text searchable copies that a folder search or index can read: what an index takes from a PDF, the steps, and how to keep it current.

To convert PDF files so they are text searchable from the folder, run OCR on every image-only file, save each result as a searchable PDF under the same folder structure, and then confirm that your search index is set to read the contents of PDFs. A folder search can only find words that are stored in the file as text.

That is two jobs, and it is easy to stop after the first. The conversion puts words into the files. The index decides whether those words are ever looked at. This guide takes both in turn, then covers searching a set of PDFs with no index at all, sidecar text files, and a routine that stops the folder from sliding back.

To see whether the index caught the new text, search one file for a party name, a docket number, and an amended complaint title you can see. Run the same search from the folder and open the hit. Compare it with the page image. When the file finds an exhibit label and the folder does not, the text is in the PDF and the index has not read it.

## Why can a folder search not find words inside your PDFs?

A folder search misses words inside a PDF for one of two reasons. Either the PDF holds no text, because its pages are pictures, or the PDF holds text and the search index has not read it. The fix for the first is OCR. The fix for the second is a setting.

### Telling the two causes apart

Pick one PDF that the folder search failed on and open it. Press Ctrl+F (Command+F on a Mac) and search for a word you can see on the page.

| Search inside the open file | Search from the folder | What it tells you |
| --- | --- | --- |
| Finds the word | Finds the file | Nothing is wrong with this file |
| Finds the word | Does not find the file | The text is there; the index is not reading it |
| Finds nothing | Does not find the file | The pages are pictures; the file needs OCR |
| Finds the word on some pages only | Finds the file for some words only | A mixed file; the scanned pages need OCR |

Run that test on three or four files from different sources before you decide what the folder needs. A matter folder collects PDFs from word processors, copiers, opposing counsel and court dockets, and each source behaves differently. The post on [why you can't copy text from a PDF](https://www.schemalaw.com/blog/why-cant-i-copy-text-from-a-pdf) explains the image-only case in full.

### A third, quieter cause

The text can be present and wrong. If a copier's own OCR read "Castellanos" as "Castellaaos", the folder search for the correct spelling skips that file, and nothing tells you it did. Indexes match characters. They do not look at the page picture. Treat a folder search that returns nothing as "not found by this search" until you have tested the set.

The difference between pressing Ctrl+F in one open file and querying an index across a folder is covered in the post on [scanning one PDF for a word versus scanning a whole folder](https://www.schemalaw.com/blog/scan-a-pdf-for-text#scanning-one-pdf-for-a-word-versus-scanning-a-whole-folder).

## What does a desktop or document-system index read from a PDF?

An index reads two things from a PDF: the text stored in the file and the file's properties, such as its name, dates, title and author. It does not read the picture of a page. A scanned page with no text layer contributes its file name to the index and nothing else.

### What Microsoft says about Windows

Microsoft's [Search indexing in Windows FAQ](https://support.microsoft.com/en-us/windows/experience/performance-optimization/search-indexing-in-windows) puts it this way: "By default, all the properties of your files are indexed, including file names and full file paths. For files with text, their contents are indexed to allow you to search for words within the files."

The same page says each file type has two possible settings: "either properties only, or properties and content." With properties only, in Microsoft's words, "indexing will not look at the contents of the file or make the contents searchable." So a folder of perfectly good searchable PDFs can still be invisible to a content search if the PDF file type is set to properties only on that machine.

Microsoft's developer documentation describes the mechanism. Windows Search uses components called filter handlers, which ["scan documents for text and properties"](https://learn.microsoft.com/en-us/windows/win32/search/-search-ifilter-about) and pass what they find to the index. The point for a law office is plain: the indexer takes text out of the file. If no text is in the file, the filter has nothing to hand over.

### Checking the Windows setting

The FAQ gives the path. In Windows Settings, select Privacy & security, choose Search, select Customize search locations, and in the Indexing Options dialog select Advanced and then the File Types tab. Find `pdf` in the list and look at the option under "How should this file be indexed?". Also confirm that the matter folder is among the indexed locations, which you change with Modify in the same dialog. On a firm-managed computer these settings may belong to your IT administrator.

### Other systems

A Mac, a network file server, a cloud drive and a document management system each have their own indexer with its own rules about which file types and locations it reads. The documentation for your system is the only reliable statement of what it does with a PDF. The test is the same everywhere: put one known searchable PDF with an unusual word in the folder, wait for the index to update, and search for the word from outside the file.

### What file properties add

Properties are searchable even when page text is not, which makes them a cheap safety net. A file named `2024-03-18 Letter Ortega to Brandt re site access.pdf` can be found by date, sender, recipient and subject before any OCR has run. Adobe's guidance on [creating PDF indexes](https://helpx.adobe.com/acrobat/using/creating-pdf-indexes.html) makes the same point about the title, subject and keywords fields in document properties, and advises using each field consistently across a collection.

## Converting a folder's PDFs so a text search reaches every file

Converting a folder means finding the PDFs that have no usable text, running OCR on those, saving each searchable PDF in the place the index looks, and testing from the folder. Work on copies, and keep the files as received somewhere the conversion never touches.

1. **Copy the originals out.** Duplicate the folder, with its subfolders, to a location named so nobody works in it, such as `Brandt v Ortega - AS RECEIVED`.
2. **List what needs converting.** Sort the PDFs into three piles: text throughout, no text, and mixed. Opening each file and searching for a visible word is slow but certain. For a large set, the post on [bulk OCR for a folder of PDFs](https://www.schemalaw.com/blog/bulk-ocr-pdf) covers the inventory and how to convert the whole folder in one run.
3. **Check each file against your OCR tool's limits.** Schema Law takes PDF, PNG and JPEG uploads up to 20 MB. The page cap per document is 3 on the Free plan, 250 on Solo and 1,000 on Team, and every plan lets you upload several files at a time. The [pricing page](https://www.schemalaw.com/pricing) has the monthly page allowances.
4. **Run OCR and export a searchable PDF.** The searchable PDF export is your original file with a hidden text layer, which gives an index text to read while the page still looks like the scan. Every plan includes it, the Free plan among them, so you can prove the approach on a short document before committing a folder to it.
5. **Save the result where the index looks.** Keep the same subfolder and the same base file name. Decide once whether the searchable copy replaces the working copy or sits beside it with a suffix such as `_ocr`, and apply that decision to every file. Two copies in an indexed folder means every hit appears twice.
6. **Let the index catch up, then test from the folder.** Search for a word that appears only in a file you converted. If the file comes back, the chain works from end to end.
7. **Test the hard cases.** Search for a surname, a dollar amount and a date from files that were faint or faxed. These are the searches that matter in a matter, and the ones a weak text layer fails.

### A worked example

Dana, a paralegal, has 140 PDFs in the correspondence folder for a construction dispute. A folder search for "retainage" returns 9 files. She knows the word runs through the pay applications. She opens one pay application, presses Ctrl+F, and gets nothing: image-only. She opens a letter the search did find and a letter it did not. The second letter has selectable text, so the index setting is also in play.

She fixes the setting first, since it costs nothing, and the count rises to 31. Then she converts the 96 files that have no text, saves them over the working copies, and waits. The count reaches 58. She spot-checks five pay applications by searching for the retainage figure on each. One fails because the figure sits under a "PAID" stamp. She notes that file in the matter log as unreliable for search and moves on.

Neither fix alone would have got her there. The setting found text that already existed. The conversion created text that did not.

### Files that are not PDFs

A matter folder also holds emails, Word files, spreadsheets, TIFF images and photos. Each has its own route to being found by a content search, and the post on [creating searchable documents from mixed file types](https://www.schemalaw.com/blog/create-searchable-documents) covers them.

## Searching many text-searchable PDFs at once from a PDF reader

Some PDF programs can search every PDF in a folder directly, reading each file in turn, with no operating-system index involved. It is slower than an indexed search and has one advantage: it reads the files as they are at that moment, so a stale or misconfigured index cannot hide a file from you.

Adobe documents this for Acrobat in its help page on [searching PDFs](https://helpx.adobe.com/acrobat/using/searching-pdfs.html). From the Find toolbar you open Advanced Search, enter the word or phrase, select "All PDF Documents in", and browse to the folder. Adobe states that Acrobat "searches the PDF body text, layers, form fields, and digital signatures", and that bookmarks and comments can be included. Results are listed under the name of each document, and the page says they can be saved as a PDF or CSV file, which gives you a record of what a search returned on a given day.

Three limits from the same page are worth knowing before you rely on it:

- **Encrypted files are skipped.** Adobe says documents encrypted for security cannot be searched as part of a multiple-document search, and that you open those and search them one at a time. If your folder holds password-protected productions, a multi-file search is silent about them.
- **Some match options exist only for multi-file searches.** Matching all of the words in any order, and Boolean queries, are described as available only when searching multiple PDFs or PDF indexes.
- **It still reads text, not pictures.** An image-only PDF in the folder produces no hits, exactly as it does in a folder search.

Acrobat Pro can also build its own index of a folder, which Adobe calls a catalog. The [PDF index documentation](https://helpx.adobe.com/acrobat/using/creating-pdf-indexes.html) says of the files you put in: "For scanned documents, ensure that the text is searchable." Every search method on this page rests on that same condition.

If you use a different PDF program, read its documentation for a folder or multi-document search and for how it treats protected files. Then run the same known-word test.

## Do you need a separate text file beside each PDF for search?

No, not for a desktop folder search. A searchable PDF carries its text inside the file, and an index that reads PDF contents needs nothing else. A separate text file per document, sometimes called a sidecar, is for systems that load text and images separately, or for indexes that cannot read PDF contents at all.

Consider a sidecar in these situations:

- **A review platform or protocol asks for it.** A production protocol can call for page images with a text file per document. The post on [TIFF versus PDF for document production](https://www.schemalaw.com/blog/tiff-vs-pdf-for-document-production) covers that format choice. Follow the specification you were given.
- **Your indexer reads `.txt` but not PDF contents,** and you cannot change it. A text file with the same base name, in the same folder, gives the index something to match. A hit on `Ortega letter 2024-03-18.txt` leads you to the PDF beside it.
- **You want to run text tools over the set.** Comparing, counting or grepping is simpler on plain text than on PDFs.

A sidecar has costs. It doubles the file count, it can drift out of step if the PDF is replaced, and a hit lands you in a text file with no page picture. If you make them, export plain text from the same OCR run that produced the searchable PDF so the two agree, and name them identically apart from the extension. In Schema Law, plain text and searchable PDF are both export formats on every plan, listed with the other [export formats on the homepage](https://www.schemalaw.com/#formats).

## Keeping new PDFs text searchable as they arrive in the folder

A folder stays searchable only if every new PDF is tested on the way in. Image-only files keep arriving from scanners, clients and other parties, and each one that is filed unconverted is a document the next folder search will not see.

A routine that holds up:

- **One intake point.** New PDFs land in an `_inbox` subfolder, never straight into the matter folders.
- **A test at the door.** Open the file, search for a word visible on a late page. Found means file it. Not found means convert first.
- **Convert before filing, not after.** A file that is filed first, with conversion left for later, is one more image-only file in the folder until someone goes back to it.
- **Name by rule.** Date, type, parties and subject in the file name, in the same order every time, so the properties search works even where the text layer is weak.
- **A monthly canary search.** Keep a short list of words you know appear in specific files. Run them from the folder once a month. A missing hit means the index, the settings or a file has changed.
- **Check after a machine or system change.** A new computer, a migrated drive or a new document system starts with its own index settings. Run the canary list on day one.

Write the routine down and name who owns it. The test at intake is the step to guard, because it is the one that gets skipped when nobody in particular is responsible for it. For checking a set that is already in place, the guide on [how to check if a PDF is searchable](https://www.schemalaw.com/blog/how-to-check-if-a-pdf-is-searchable) covers testing page by page and file by file.

## 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)