Troubleshooting guide
Paperless-ngx OCR Not Working: Empty Text Layer After Import
Category: Document Workflow
You imported a PDF through your Docker watched volume, but the document opens with no selectable text and searches return nothing, so you are stuck at the ocr produces an empty document step.
Confirm the file actually needs OCR
First, check whether the PDF already has a text layer. Run pdftotext -layout file.pdf - | head on the host. If it prints readable words, OCR is not the problem; the import or consumption stage is. If it prints nothing, continue.
Check the OCR language and image quality
Inside the container, run tesseract --list-langs. If your document’s language is missing, install the matching language pack in your Dockerfile or image. Then inspect the scan: a low-resolution or skewed image can make OCR return empty output. Try pdftoppm -r 300 -png file.pdf page and run tesseract page-1.png stdout to see whether text appears at 300 DPI.
Verify the watched volume and permissions
The consumer must read the file. Check docker logs paperless-ngx for permission errors or skipped files. Confirm the watched folder is mounted read-write and that the container user can read the PDF. If the file is moved before consumption finishes, OCR may run on a partial file.
Verification test
Place a single clean, 300 DPI scanned PDF into the watched volume. Wait for the consumer to finish, then open the document and select text. A successful result shows selectable words and search hits. If the text layer is still empty, the issue is in OCR configuration, not the import path.
Rollback and recovery
Before changing language packs or image settings, copy your compose file and note the current image tag. If a change breaks consumption, restore the previous compose file and restart the stack. Re-upload the test PDF to confirm recovery.
If these checks do not resolve your paperless ngx ocr not working case,

Ask for help with your logs and test PDF details.