Ready signal
Ready for OCR attempt
The file has no usable text layer, scan-quality signals are acceptable, rotation/language concerns are known, and you have a checklist for reviewing names, numbers, and dates.
Quality-gated workflow
Scan Rescue and OCR Readiness
Use Scan Rescue when text extraction failed, a PDF appears scanned, or an image contains text that needs review. It helps decide whether OCR is worth running and how to verify output; it does not repair scans or promise recognition accuracy.
Start here
Choose the next recovery path
Text extraction returned empty Check text-layer, scan, rotation, and image-density signals before launching OCR. Not sure if OCR is needed Use the OCR decision guide before spending another recognition run. OCR output seems wrong Review quality limits, confidence, rotation, and language expectations. Editable Word is the goal Use PDF-to-Word readiness separately from OCR text output.
Preflight result
Scan rescue readiness
Use these signals before running server-backed OCR. Ready means the file is a reasonable OCR candidate with a review plan, not that recognition will be perfect.
Needs review
Use text extraction first
If the PDF has selectable text or digital text-layer signals, PDF to Text is usually cleaner and cheaper than OCR.
Needs review
Needs scan cleanup review
Low contrast, rotation, tiny text, blank pages, mixed page sizes, or photo glare should be reviewed before spending a server OCR job.
Needs review
Needs output verification
OCR output should be checked against the source for names, totals, punctuation, line breaks, and table structure before reuse.
Task groups
Start with the part of the workflow that is failing
Classify the source first
Check for a PDF text layer, scanned-page signals, existing extractable text, page rotation, and image-density clues before launching recognition on a file that may already be digital.
Inspect scan and image quality
Review scan quality, image contrast, blankness, dimensions, file format, metadata, and language-readiness signals before accepting OCR output as usable.
Run bounded OCR jobs
Use temporary OCR jobs for scanned PDFs or images only after readiness checks make likely quality limits clear, then compare the output before downstream reuse.
Route editable-document expectations
If the destination is editable Word, check PDF-to-Word readiness separately because OCR text and editable DOCX conversion are different workflows.
Use guides before retrying
When the output is empty, noisy, or misread, use the OCR path chooser and scanned-document guides before launching another conversion or sending the result onward.