Make scanned PDFs searchable with OCR Pro

Scanned or photographed PDFs are pictures of pages — they have no text layer, so FileDeck’s full-text search cannot read them. Enable OCR under Documents → Settings → Scanned OCR and FileDeck sends each scanned PDF you publish to the FileDeck OCR service, which extracts the text and returns it, making the document searchable like any other. Scans already in your library when you enable it, and scans brought in by an import, are sent when you press Send scanned PDF(s) for OCR on the Search Index tab.

FileDeck Pro — this feature is included in all Pro plans. See pricing

How does OCR handle my files?

FileDeck is built to keep your documents on your own server: normal full-text search reads the text layer of your PDFs locally, and nothing leaves your site. OCR is the one exception.

• Apart from the AI & Governance features, which send text to the AI provider you connect with your own key, OCR is the one FileDeck feature that sends document contents off your server. It sends only PDFs FileDeck could not read a text layer from on your server (normally scans), and only while OCR is enabled: each scan as you publish it or give it to a document as a new file, and the scans already in your library, including any an import brought in, when you press Send. Documents it could read are not sent.
• Files are processed and immediately discarded — the OCR service never stores them.
• Each file is sent with its file name and your site’s license install ID, which the service uses to confirm your license. Like any request WordPress makes, it also carries your site’s address.
• OCR is off until you enable it, so nothing is sent without your explicit opt-in.
• Access rules are not a filter here: a scanned PDF in a restricted or password-protected category is sent for OCR exactly like a public one.
• If your server is missing a PHP extension FileDeck needs to read PDFs (zlib, iconv or mbstring), no PDF is sent: FileDeck can’t tell a scan from a PDF with a text layer there. The Search Index tab names the missing extensions.

How do I enable OCR?

1. Go to Documents → Settings → Scanned OCR.
2. Tick Enable OCR.
3. Save your changes.

There is no key to enter. OCR is included with your Pro license, and FileDeck uses your activated license to connect to the OCR service.

How do I run OCR on my documents?

Once OCR is enabled, each PDF FileDeck cannot read a text layer from is sent in the background when you publish it, including a scan that replaces a document’s file. There is no button to press. A draft is sent when you publish it.

Scans that were already in your library when you enabled OCR wait for you, and so do scans added on Documents → Import & Export — bulk add (including files you drop there), CSV import and migrations — including imported drafts you publish later. Turning OCR on, or importing a library, never sends an existing collection by itself:

1. Go to Documents → Settings → Search Index.
2. Click Send N scanned PDF(s) for OCR. FileDeck works through the PDFs it could not read a text layer from, one at a time in the background.
3. Click Rebuild search index after a bulk import or whenever results look out of date, so newly extracted text is included.

Once a scanned PDF has been through OCR, it matches in instant search just like a native PDF — visitors don’t see any difference.

How long does OCR take?

OCR runs in the background, one document at a time, using WordPress’s scheduled tasks. A short scan takes seconds; a 40-page scan can take a few minutes. The Search Index tab shows how many documents are still queued. If you publish a very large batch of scans at once, they join the queue as it drains rather than all at once.

WordPress runs scheduled tasks when someone visits the site, so on a site with very little traffic the queue moves only as visits come in. If your host has turned off WordPress’s built-in scheduler (DISABLE_WP_CRON), make sure its replacement cron job is running.

If a document fails — the OCR service didn’t answer in time or was briefly unavailable — FileDeck retries it automatically, waiting longer between each attempt. After five failed attempts, or straight away if the service refuses the key or the file, it stops. The Search Index tab then says why, and clicking Send scanned PDF(s) for OCR tries again.

FAQ

Does OCR send all my documents off my server?

No. Only PDFs FileDeck could not read a text layer from on your server are sent to the OCR service, and only while OCR is enabled. Content search reads native PDFs, Office and OpenDocument files, and text files on your own server; the OCR service does not see them.

Does turning OCR on send the scans already in my library?

No. Enabling OCR sends only the scans you publish, or give a document as its new file, from then on. Scans that were already in your library wait until you press Send scanned PDF(s) for OCR on the Search Index tab, even when the search index is rebuilt.

Does importing a library send its scans?

No. Scans brought in by an import wait for Send, even with OCR on: a one-click or CSV import, bulk add (including new files you drop into it), or attaching migrated files from Library Health. That includes imported drafts you publish later. A new file you give an imported document afterwards is sent like any other.

How do I stop new scans being sent?

Untick Enable OCR and save. Nothing more is sent while it is off, including scans that were still queued. If you turn it back on, press Send scanned PDF(s) for OCR to finish those.

Are my files stored by the OCR service?

No. Files are processed and immediately discarded. They are never stored.

Is there a limit on how much I can OCR?

The managed OCR service is shared, so there are fair-use limits: up to 100 scanned documents a day for your site, and 2,000 pages a day in total. Only the first 40 pages of each scan are read. If you send a large backlog of scanned files, they’re processed a batch at a time and the remainder continue automatically over the following days. (Running your own OCR service removes the service’s page limit; the 100-a-day pace can be changed with the filedeck_ocr_daily_cap filter.)

Do I need an OCR key?

No. Your activated Pro license is all OCR needs. If you were sent a site key before OCR was included with licenses, it keeps working; you’ll find it under Use your own OCR service instead on the same screen. To switch to your license, empty the Service URL, tick Remove the saved key, and save.

Can I use my own OCR service?

Yes. Under Use your own OCR service instead, enter your service’s URL and site key. The key is stored securely and not shown again after you save.

Enter just the service’s address, such as https://ocr.example.com. If you copy it from an email, your email client may have turned it into a long redirect link (starting with www.google.com/url?q= or ending in safelinks.protection.outlook.com); FileDeck saves the address inside such a link and tells you it did. After saving, click Test connection on the same screen to check that the address reaches an OCR service.

Why is a scanned PDF still not matching in search?

Open Documents → Settings → Search Index. If the document is still queued, give it time (see How long does OCR take?). If the tab says OCR stopped on it, the reason is shown there; fix that and click Send scanned PDF(s) for OCR again. Otherwise, check that OCR is enabled under Documents → Settings → Scanned OCR and that your Pro license is activated, then rebuild the index from Documents → Settings → Search Index. See Troubleshooting search for more.