# Searchable PDF for Court Filing: What Courts Require

- Author: Schema Law (https://www.schemalaw.com/)
- Published: 2024-05-26T12:14:45Z
- Updated: 2026-10-11T19:22:20Z
- Category: Litigation support
- Canonical: https://www.schemalaw.com/blog/searchable-pdf-for-court-filing

> Do you need a searchable PDF for a court filing? Rules from named courts compared, plus how to prepare, size and check a scanned exhibit first.

Whether you need a searchable PDF for a court filing depends on the court. Some courts require every filed PDF to be text-searchable, some only where feasible, and at least one tells filers not to OCR scans. To produce one, convert your own drafting directly to PDF, and scan paper exhibits with OCR switched on.

This post compares what several named courts say, in their own words, and then walks through preparing a scanned exhibit so that it meets a searchability rule and a size limit at once. It explains document handling and is not legal advice. Every example below was read on the court's own website, and each of them can change, so treat them as illustrations and confirm the current rule where you practice.

## Do courts require text-searchable PDFs?

Many do, some do not, and the details differ. There is no single national rule on searchability. Each court sets its own requirements through local rules, administrative orders and filing guides, and those documents disagree with each other on scanned papers in particular.

Electronic filing itself is close to universal for lawyers in federal court. [Rule 5(d)(3)(A) of the Federal Rules of Civil Procedure](https://www.law.cornell.edu/rules/frcp/rule_5) says a person represented by an attorney "must file electronically" unless the court allows otherwise for good cause or a local rule allows or requires paper. The rule says nothing about whether the PDF must be searchable. That is left to each court.

Here is what five courts say.

| Court | What its own material says | Where to read it |
| --- | --- | --- |
| U.S. Court of Appeals for the Second Circuit | A PDF filed in CM/ECF "must be text-searchable", and the court "will return a PDF that is not text-searchable to a filer for resubmission" | [Second Circuit guidance on working with PDFs](https://coop.ca2.uscourts.gov/clerk/case_filing/electronic_filing/how_to_use_cmecf/working_with_pdfs.html) |
| U.S. Bankruptcy Court, Southern District of Florida | "All PDF documents created for filing in CM/ECF", proposed orders included, "must be saved in a text-searchable format" | [Administrative Order 2022-03](https://www.flsb.uscourts.gov/sites/flsb/files/documents/general-orders/AO_2022-03_Amendment_of_Local_Rule_5005-1_to_Reflect_Requirement_for_Submission_of_Electronically_Filed_Documents_in_Text-Searchable_PDF_Format.pdf) |
| California state trial courts | An electronically filed document "must be text searchable when technologically feasible without impairment of the document's image" | [California Rules of Court, rule 2.256(b)(3)](https://courts.ca.gov/cms/rules/index/two/rule2_256) |
| U.S. District Court, District of Arizona | For scanned documents: "Do not use Textbridge or OCR software" | [District of Arizona scanning guidelines](https://www.azd.uscourts.gov/content/what-are-guidelines-scanning-documents) |
| U.S. District Court, Western District of Washington | Its file size page sets scanning requirements of 300 dpi and black and white, and does not mention OCR | [Western District of Washington file size guidance](https://www.wawd.uscourts.gov/node/581) |

Three patterns come out of that table.

**Strict courts treat an image-only PDF as a defective filing.** The Second Circuit page calls such a file "non-conforming". That page shows a last-modified date in 2010, so read it as the court's posted guidance and check the Second Circuit's current local rules before you file. A returned filing close to a deadline is a real problem, so in a court like that the searchability check belongs on your pre-filing list.

**Qualified rules leave room for bad originals.** California's "when technologically feasible" wording accepts that some pages cannot be made searchable without harm to the image.

**Some guidance points the other way.** Arizona's scanning guideline tells filers not to use OCR software on scans, and says in the same place that every document not scanned from paper must be converted directly from a word processor and be text searchable. The page is undated, so confirm it against that court's current rules. A habit learned in one court can be wrong in the next.

Where do you find your court's position? Look in this order: the local rules, any administrative or general order on electronic filing, the clerk's CM/ECF or e-filing user manual, the court website's FAQ, and your judge's individual practices. If they conflict or are silent, call the clerk's office.

## What does a court mean by text-searchable?

A court that requires a text-searchable PDF means a file in which the words can be found with a search and selected with a cursor. Courts recognize two ways to get there. Converting a word-processing file straight to PDF is one. Scanning paper with optical character recognition switched on is the other.

The Southern District of Florida bankruptcy court's order is unusually explicit, and worth reading as a definition. It says a document "is text-searchable when converted to PDF format directly from a word processing program using Adobe Acrobat, Windows PDF or other similar software." For paper, it says "the scanning software must have Optical Character Recognition (OCR) functionality enabled."

Run OCR when the court you are filing in asks for a text-searchable PDF, and keep the page image under the text. A federal court may put that request in its CM/ECF manual. A state court may put it on a different page. Read the page for your judge and for the clerk of court before you file the motion or the exhibit. Then search for a word you can see.

California's rule comes with an advisory committee comment that sets the level of effort. "Technologically feasible", the comment says, "does not require more than the application of standard, commercially available optical character recognition (OCR) software." In other words, the rule asks you to run ordinary OCR. It does not ask you to retype a page the software cannot read.

Several practical points follow from definitions like these.

- **Your own drafting should never be scanned.** A brief, motion or declaration that you wrote exists as a word-processing file. Save or print it to PDF directly. Printing it on paper, signing it and scanning the whole thing turns good text into a picture.
- **A signature does not force a scan of everything.** If one page needs an ink signature, scan that page and combine it with the directly converted pages. Check your court's rules on electronic signatures first, since an "/s/" signature may make even that unnecessary.
- **Searchable means the text layer, not the look.** A scan can be perfectly legible on screen and still fail the requirement. The court's concern is whether the judge and clerks can search and copy.
- **Page numbers count too.** The Second Circuit's guidance speaks of paginating a PDF so that page numbers are "properly text-searchable". Bookmarks, pagination and combining files are separate requirements in many courts. Read the whole section of the rule, not only the searchability sentence.
- **Nobody defines accuracy.** None of the rules quoted here sets a standard for how correct the hidden text must be. That does not make accuracy unimportant. A judge who searches your exhibit for a name and gets no hit has been let down by your filing.

How a text layer sits behind a page image, and how it is added, is explained in the guide on [how to make a scanned PDF searchable](https://www.schemalaw.com/blog/how-to-make-a-scanned-pdf-searchable).

## Preparing a scanned exhibit for electronic filing

Prepare a scanned exhibit by starting with the best source you can get, scanning at the settings your court asks for, adding a text layer if the court wants one, and then checking size, searchability and redactions. Exhibits are where searchability rules bite, because they are the part of a filing you did not write.

1. **Look for an electronic original.** An email, a spreadsheet or a contract that exists as a file should be converted directly to PDF. Scan only what exists only on paper.
2. **Read the court's scanning settings.** Courts that publish settings tend to ask for modest ones. The Western District of Washington's page calls for "300 dpi and black and white unless color is integral to the document." The [Northern District of Texas bankruptcy court's scanning guidelines](https://www.txnb.uscourts.gov/content/scanning-guidelines) say resolution "should be set to 300 dpi (DO NOT exceed 300 dpi)" and image type to black and white. General advice on resolution and color for OCR is in the guide to [digitizing law firm paper files](https://www.schemalaw.com/blog/how-to-digitize-law-firm-paper-files#scanner-settings-that-make-ocr-work).
3. **Scan each exhibit as its own file.** Separate files are easier to label, to size and to replace when one turns out badly.
4. **Add the text layer if your court requires or allows it.** Either switch on OCR in the scanning software or run the scanned PDF through an OCR tool and export a searchable PDF. In Schema Law that means uploading the PDF and exporting it as a searchable PDF. All plans, the free one included, offer that export, and the [pricing page](https://www.schemalaw.com/pricing) shows the page limits for each.
5. **Do not let OCR change the page.** The exhibit must look exactly as the original did. A searchable PDF that keeps the scanned image and hides the text behind it does that. A conversion that re-flows the document into fresh type does not, and is the wrong thing to file as a copy of an exhibit.
6. **Redact before you add text, then check again after.** A text layer made before redaction can still hold the hidden words.
7. **Assemble, label and bookmark as the rules require.** Follow the court's instructions on combining exhibits, cover pages and attachment descriptions.

### What kind of text layer are you filing?

Text layers are not all built the same way, and it is worth knowing which kind your tool makes. Many tools place each recognized word directly over its picture, so a search highlights the word where it appears.

Schema Law's searchable PDF export works differently. When you upload a PDF, the export is that same file with an invisible layer added to each page that holds the page's text in reading order. The layer is not positioned word by word. A search finds the page that contains the term, and the highlight may not rest on the matching words. Selecting text with the cursor will also not track the visible lines closely.

None of the rules quoted in this post says where on the page the hidden text must sit. Even so, test one exhibit in a standard PDF reader and decide whether the result suits your court and your judge before you rely on it for a filing. If you need word-level highlighting, use a tool that places text under each word. Schema Law's stronger uses are getting the content out as Word, Excel or text, and making a set searchable for your own review, as in the guide to [OCR for trial exhibits](https://www.schemalaw.com/blog/ocr-trial-exhibits).

## File size limits and scanned attachments

Scanned attachments are the usual reason a filing hits a size limit, because page images are far larger than text. Each court sets its own cap per file. Know yours before you scan, pick settings that keep files small, and split an oversized exhibit into labeled parts.

Limits from three courts show how wide the range is.

| Court | Limit per PDF | What it tells you to do when over |
| --- | --- | --- |
| [U.S. Bankruptcy Court, Northern District of Texas](https://www.txnb.uscourts.gov/content/scanning-guidelines) | 25 MB | Check scanner settings, then split into multiple PDFs |
| [U.S. District Court, District of Arizona](https://www.azd.uscourts.gov/content/what-are-guidelines-scanning-documents) | 30 MB | Extract pages to a separate file |
| [U.S. District Court, Western District of Washington](https://www.wawd.uscourts.gov/node/581) | 100 MB | Divide it into parts under the limit and label each one, such as Exhibit 1, Part 1 and Exhibit 1, Part 2 |

Each court name in the table links to the page the figure comes from. Arizona's page also estimates how many pages fit under its limit, and the estimate is much lower for tables and graphics than for plain text. A table-heavy exhibit reaches a limit far sooner than a typed letter.

What controls the size of a scan:

- **Color mode.** Black and white is smallest by a wide margin. Grayscale is several times larger and color larger again. Use color only where it carries meaning, such as a photograph or a highlighted chart.
- **Resolution.** Higher resolution multiplies file size. The Western District of Washington's page warns that scanning above its 300 dpi requirement "can cause lagging/timeout issues while uploading."
- **Page content.** Dense tables, photographs and dirty backgrounds compress poorly.
- **The text layer.** Hidden text adds very little. Characters are tiny next to images. Making a file searchable is rarely what pushes it over.

When a file is still too large, split it. Cut at a natural break, such as between sub-exhibits or at a tab, and label the parts the way your court's guidance says. Check the page total of the parts against the original, and make sure the attachment descriptions tell the reader what is in each part.

A common mistake is fixing size with heavy compression or by dropping to a very low resolution. That can make small print unreadable to a person, and it wrecks OCR for anyone who later needs to search the page. Split before you degrade.

One more number matters if you use an online tool to add the text layer. The tool has its own upload limit, which may be lower than the court's. Schema Law accepts files of up to 20 MB, below all three court limits in the table, so a large exhibit may need splitting before conversion. The guide to [batch OCR for discovery documents](https://www.schemalaw.com/blog/batch-ocr-discovery-documents#reassembling-exports-into-one-set) covers putting parts back together in order.

## Should you file the scan or a converted document?

File the form that is both faithful and searchable. For anything you wrote, that is a PDF made directly from the word-processing file. For a paper exhibit, it is the scanned image, with a text layer if your court wants one. A re-typed conversion of an exhibit is a different document and should not stand in for the original.

| Document | What to file | Why |
| --- | --- | --- |
| Brief, motion, declaration you drafted | PDF converted directly from the word processor | Exact text, small file, meets every searchability rule quoted here |
| Signed declaration from a client, received as a scan | The scan, with OCR if required; or a directly converted PDF with a conforming electronic signature if your court allows | The signature page must be what was signed |
| Paper exhibit: contract, letter, record | Scanned image, plus text layer where required or allowed | The court needs to see the document as it exists |
| Exhibit you hold as an electronic file | PDF made from the file, if that is a fair copy of the exhibit | Cleaner and smaller than print-and-scan |
| Exhibit too poor for OCR to read | The best scan you can get, searchable where feasible | Rules such as California's qualify the requirement by feasibility |
| A Word conversion of a scanned exhibit | Do not file it as the exhibit | It is your software's reading, with possible errors and a changed layout |

That last row is the mistake to avoid. A Word export of a scanned contract is useful for drafting and for quoting, and the guide to [OCR for court filings](https://www.schemalaw.com/blog/ocr-court-filings) shows how to make and check one. Filed as "Exhibit A", it misrepresents what the contract looks like and may contain words the original does not. If a readable version would help the court, for example with a barely legible handwritten letter, one approach is to file the original image and offer a transcription as a separate, clearly labeled document. Check whether your court permits that and what it requires.

A second question is whether to OCR at all in a court that discourages it. Follow the court's instruction for the filed copy. Nothing stops you from keeping a searchable copy in your own file for preparation.

A third is the mixed document. A filing that combines converted pages and scanned pages will be searchable in places. If your court requires searchability throughout, run OCR on the scanned pages before you combine them, then test the result page by page.

## Checking a PDF before you file it

Check four things in the final PDF, in the reader you would expect the court to use: that text can be found on every page, that redacted material is gone from the text layer as well as the image, that the file is within the size limit, and that the pages are complete and in order.

**1. Searchability, page by page.**

- Open the exact file you will upload, not an earlier draft.
- Search for a word visible on the first page, one in the middle and one on the last page.
- For a combined filing, test at least one page of every exhibit. A single image-only attachment is what triggers a rejection in a strict court.
- Select a line and paste it into a blank document. Check that the pasted words are real text and not scrambled characters.

**2. Redactions.**

This is the check with the most at stake. [Rule 5.2(a) of the Federal Rules of Civil Procedure](https://www.law.cornell.edu/rules/frcp/rule_5.2) limits what a filing may show of social-security numbers, taxpayer-identification numbers, birth dates, minors' names and financial-account numbers. A black box drawn over a number on the page image does not remove that number from a text layer made before the box was drawn. Anyone can then find it with a search or copy it out.

- Redact first, flatten the page to an image, and only then run OCR. The recognition can read only what is still visible.
- After OCR, search the final file for the redacted items themselves: the full account number, the full date of birth, the minor's name.
- Select all the text on a redacted page, paste it elsewhere and read it.

**3. Size and format.**

- Check the file size against your court's per-file limit.
- Confirm the file is a PDF of the kind the court accepts. Some courts have further requirements, such as flattening fillable forms, which the Second Circuit's guidance mentions.
- Make sure no password or security setting is on the file.

**4. Completeness.**

- Count the pages against the source.
- Flip through for rotated, blank or duplicated pages.
- Check that bookmarks, exhibit labels and page numbers match the index or table of contents.

A short version to keep by the scanner:

| Check | Pass when |
| --- | --- |
| Search first, middle and last page | Each search finds the word |
| Paste a selected line | Pasted text is readable |
| Search for each redacted item | No results |
| File size | Under the court's limit |
| Page count | Matches the source |
| Court's own rule | Read this month, not remembered |

If the document will not pass the search test, the cause is usually an image-only page, and 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) shows how to confirm that. For a confidential exhibit, also consider where the file goes when you use an online tool to add text. Schema Law's [security commitments](https://www.schemalaw.com/#security) state how uploads are treated, and you can delete a document once the filing is done.

## 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)