NetDocuments OCR: What the Platform Says It Does

NetDocuments OCR exists, but it is not on for every repository. NetDocuments describes it as a service a customer subscribes to and an administrator configures cabinet by cabinet. Where it is running, scanned PDFs and image files you save are converted to searchable PDFs. Where it is not, a scan is found only by its profile data.

Everything below about the platform comes from NetDocuments' own support articles, opened on October 8, 2026 and linked where they are used. Your firm's configuration decides which parts apply to you, so treat this as a list of things to verify, not a description of your repository.

Does NetDocuments OCR scanned documents?

Yes, when your firm has the service and has switched it on for the cabinet. The NetDocuments OCR FAQ describes a background process that monitors files users import and converts image-based documents into text-searchable ones, so the platform's full-text search can find them.

The same FAQ says there are "two services that customers can subscribe to and enable", and that most customers need both:

• Backlog/Bulk Upload crawls the existing repository and processes the documents that need it.
• Active Monitoring processes documents and images as they are added.

The word to notice is subscribe. NetDocuments' marketing page on legal file search presents OCR as the way to find non-searchable images and files in the system, but the support articles make plain that it has to be set up. If nobody at your firm remembers doing that, do not assume it was done.

What NetDocuments' documentation says its OCR covers

NetDocuments' support pages say the service reads image-based PDFs and TIFF, BMP, PNG and JPG files, and they list EML and MSG email files among the supported formats. The result is a searchable PDF saved back to the same document, either as a new version or over the current one. Nothing is installed on your computers.

The details come from the FAQ and from the article on NetDocuments OCR administration:

Two conditions in those articles are easy to miss. The administrator picks a backlog scope, meaning how far back the service reads, and the FAQ suggests that five years is often enough. Anything older than the chosen date stays as it was. The administrator also sets a character threshold per page, 150 by default, that tells the service when a page already has enough text for OCR not to take place.

Why does a scanned PDF not appear in NetDocuments full-text search?

Because the search index holds no words for it. NetDocuments' article on the rules of searching says that if a PDF was created from a scanned or faxed image and was not processed through OCR, only its profile data can be searched.

Work through the causes in this order. Each one comes from the support articles cited above.

1. You just saved it. NetDocuments uses one index for metadata and another for full text, and says full-text indexing may take longer. Search by name or profile field first, then try the content again later.
2. OCR is not subscribed or not configured. The service is set up from the OCR Dashboard under Repository Administration.
3. The cabinet or file type is switched off. OCR is applied per cabinet, and the administrator selects which file types to include.
4. The document predates the backlog scope.
5. The document was checked out or locked when the service reached it. It is tried again the next time the document is full-text indexed.
6. The file is password protected, carries a digital certificate, or is corrupted. The FAQ lists each of these as a reason a document is not processed.
7. There is too little text. The dashboard reports "Not enough words detected" when fewer than three words are extracted.
8. The page is too poor to read. The FAQ says that if a person would reasonably struggle to read a document, it will generally not OCR at all, or not very well.

To see whether a particular document was processed, open its history. NetDocuments says a processed document shows a change made by a system user called "NetDocumentsOCR Agent", and the version description reads "OCR'd from version" followed by the version number.

Remember what kind of problem this is. The scan in the repository is an image-only file, and nothing about saving it to a document management system changes that. The neighbouring post on SharePoint OCR describes the same gap in Microsoft's platform.

Questions for your NetDocuments administrator about OCR

Ask whether the firm subscribes to NetDocuments OCR, which of the two services it has, and how the settings were chosen. Six short answers will tell you whether a scan you save today will be searchable tomorrow, and whether the scans from past years ever were.

• Do we subscribe to Backlog/Bulk Upload, Active Monitoring, or both?
• Which cabinets have OCR turned on, and which file types are included?
• What is the backlog scope date? Are closed matters older than that still image-only?
• Is the save setting "Save as new version" or "Overwrite current version"? This decides what the official version of a scanned exhibit is after processing.
• Which languages are selected? A firm with Spanish-language client records needs Spanish on the list of three.
• What does the OCR Dashboard show under exceptions, and who reviews it?

The save setting deserves a second look in litigation work. If the official version of a produced document becomes the OCR output, be sure everyone knows where the scan as received is kept. Under the overwrite setting, NetDocuments says it is stored as an attachment to the new version.

Once indexing has had time to run, look in the cabinet for a sentence you remember from the complaint, the sticker on an exhibit, and the witness name on a deposition cover. Do it again for a pleading you know is in the discovery set. When a visible word on the scan is missing from the results, note the document and ask whether its text reached the index.

Making a scan searchable before you save it to NetDocuments

If you want OCR for NetDocuments files without relying on the platform's own service, add the text before the file goes in. A PDF that already has a text layer is no longer the unprocessed scan the search rules describe, and it stays searchable after you download or send it.

This is the upstream route, and it suits three situations: your firm does not subscribe to the OCR service, the document falls outside the backlog scope, or you need the file itself to be searchable when it leaves the repository. The guide to how to make a scanned PDF searchable explains what the text layer is and how to confirm it is there. The comparison of ways to make a PDF searchable without Acrobat lists the tools that can add one.

Schema Law is one of them, and it has no connection to NetDocuments. You upload the scan in a browser, download the searchable PDF from the export formats on offer, and save that file to the workspace the way you save any other document.

Test with one file before you convert a matter's worth. Save the searchable PDF, wait for full-text indexing, and search for a phrase from the middle of a page. Then check the document history. If your firm's OCR service is also running, find out whether it leaves a file that already has text alone, which is what the per-page character threshold is for.