A PDF can contain text objects, page images, or a mixture. A scan may look like ordinary text while the file initially contains only pixels. Optical character recognition (OCR) can add or infer text, but the recognized result still needs checking.
Compare two original practice files
Download the text-based PDF and the image-only PDF. Both show the fictional phrase Cedar workshop: 16 seats. One is generated with text objects; the other contains a raster image of that text and no embedded text layer at creation.
Open both in a supported PDF viewer. Search for Cedar and try selecting a word. Record the viewer and what happens instead of assuming the image file must always be unsearchable.
Account for viewer OCR
Current Chrome documentation describes on-device OCR for scanned PDFs. A viewer with recognition can make the image-only sample selectable or searchable even though its original file contains no text objects. That is why “I can search it here” and “the file contains a portable text layer” are not identical claims.
| File or behavior | What it establishes |
|---|---|
| Text objects embedded in PDF | A text representation exists in the file |
| Image-only source | Visible words begin as pixels |
| Viewer recognizes scanned words | That viewer can provide inferred text |
| Search succeeds once | The search worked in that environment; not universal compatibility |
Rule out other causes of selection failure
In Preview, use Tools → Text Selection before trying to select text. A different tool may be active. PDF permissions or a password-protected document can also limit copying. Not every selection failure means the page is a scan.
Try a distinctive short word rather than an entire sentence. Line breaks, ligatures, and OCR errors can make a phrase search fail even when some text is available. Check whether the document uses a font encoding that produces unexpected copied characters.
Verify recognized content before relying on it
Compare names, dates, and numbers with the visible page. OCR can confuse 1 and l, drop punctuation, or reorder columns. For the sample, check both Cedar and 16 seats; a partly correct sentence can conceal an important number error.
If you need to share a reliably searchable file, use an authorized OCR or export workflow and test the saved output in the recipient’s expected viewer. Do not upload sensitive scans to an unknown conversion site. Searchability helps retrieval, but it does not by itself establish accessibility, reading order, or factual accuracy.
References & scope
These official references support the software instructions. Practice documents are original, fictional teaching examples; labels can vary by release.
- Select and copy text in a PDF in Preview on Mac
Use Text Selection and Copy for PDF text; column selection is distinct. - If you cannot select or copy text in a PDF in Preview on Mac
Selection tools and PDF permissions can prevent copying; a scan is not the only cause. - Download a file
Chrome download settings and saving edited PDF form changes.