Recover and re-use content from a damaged file
Goal
Get a usable PDF, or at least the text and pictures, out of a file that will not open, opens blank, or shows broken text.
A download was cut off at 80%. Your viewer says "the document is damaged". The client's photos and the agreement text are in there somewhere, and you cannot ask them to resend until Monday.
Starting state
- One damaged PDF (Recovery works on one file per run).
- If the file is also password-protected and you have the password, unlock it first. Recovery reports encrypted files as unrecoverable. If you do not have the password, nothing here can help.
- Keep the original; try different settings from it, not from a previous result.
Chain
Unlock (if locked) → Recovery → Extract Text / Extract Image (or Parser)
only with password repair take out what you need
The plan listed Recovery before Unlock. The chapters say the opposite is required: Recovery refuses locked files, so Unlock comes first when it applies.
Step by step
1. Unlock (only if the file asks for a password)
Type the password and press Unlock. See Unlock.
2. Recovery
Open Advanced options:
- Expected page count: type the true number if you know it. A repair with fewer pages is rejected rather than handed to you as "success".
- Minimum text coverage (%): set a high number for a text document; leave 0 for scans, which have no text and would be rejected.
- Sanitization policy: the default Full strips metadata, attachments, history and hidden content. If you want attachments, choose Keep attachments; if you want the file returned as repaired, choose None.
- Keep Include full diagnostic report (PDF X-ray) on.
Download repaired.pdf and any extra files (images, attachments, metadata, recovered text or form values). Read the X-ray: which repair stage worked, which pages are original, repaired or reconstructed, and the confidence level. See Recovery.
3. Extract Text
On repaired.pdf, choose Pure, press Extract. Why: even if page layout is imperfect, the words may be intact. If text is garbled, check the report's font repairs. See Extract Text.
4. Extract Image
Press Extract to get every embedded picture as PNG. Why: pictures often survive when pages do not. See Extract Image.
Prefer one download? Parser gives text, images, attachments, annotations, bookmarks and form data in one ZIP.
Why this order
Repair before extraction, because extractors expect a readable file. Unlock before repair because encryption blocks it. Extraction last because you are taking out of the best version you now have.
Expected result
A repaired PDF with an X-ray report, plus .txt and image files, or at least those extras if the PDF itself stays beyond repair.
Check it worked
repaired.pdfopens in your viewer and shows the expected page count.- The report lists no unexpected dropped features and a confidence level you accept.
- Reconstructed pages (marked in the report) are compared against anything you have, such as an earlier copy.
- The .txt contains the right words in a spot-check; image files open.
Variations
- Text looks garbled, or scanned: run Searchable on the repaired file to rebuild text from the page images.
- Just need it small again: repair, then Compressor; never compress a file that is not yet healthy.
- Not actually a PDF: Recovery hands you the file back under its real extension (for example a .docx or .jpg).
What can go wrong
- "No recovery stage produced an acceptable result": lower the page or text gates, read the damage notes; maybe too little is left.
- Metadata and attachments vanished: that is the Full policy; rerun with None or Keep attachments.
- A strict text-coverage gate on a scan rejects good repairs.
- Do not assume reconstructed pages equal the original; the report says which are which.
- The file stays unusable: the extras (images, text) may still be worth taking.