What this covers
- No text layer is the cause in most cases
- A permission flag is a reader honouring a request, not a lock
- Garbled copied text is a broken character map, not your fault
- Only the second of the three is fixable
You can see the words. You drag across them and nothing happens, or something happens and what lands in the clipboard is not what was on the page. There are three distinct causes and they need completely different responses.
Cause one: there is no text
By far the most common. The document is a scan - a photograph of a page - and it contains no characters at all. The letters you can see are pixels arranged in a shape your visual system recognises.
Nothing about the appearance reveals this. A scan and a generated PDF look identical on screen. The test is the attempt itself: if the cursor sweeps across as though over a photograph, or the whole page turns blue at once, there is nothing there to select.
There is no fix within the document. Only optical character recognition would create a text layer, and it would be a machine's reading rather than the original text.
Cause two: a permission flag
A PDF can carry a flag asking readers not to allow text extraction. The document is fully readable; the flag is a request, and a well-behaved reader honours it by making the text unselectable.
The distinguishing symptom is that the document opened without a password and the text simply will not select, while the page is obviously composed of real text. Zooming in confirms it: text drawn from a font stays perfectly sharp, and a scan does not.
This one is fixable. Rewriting the file clears the flag - Unlock PDF will do it, and will tell you in the result that the document was never actually encrypted, which is worth knowing.
Whether you should is a separate question
Clearing a restriction on your own document is unremarkable. Doing it to material that was restricted deliberately by someone else, in order to reuse their content, is a different act. The tool requires you to have the file; it cannot know which situation you are in, and you do.
Cause three: the text copies as nonsense
Less common and more confusing. The text selects normally, you paste it, and what appears is a run of unrelated characters or symbols.
This happens when a font is embedded with a non-standard encoding and the document does not carry a correct mapping back to Unicode. The reader draws the right glyphs, because it is following the font's internal codes, but the file does not say which actual characters those codes represent. Copying gives you the codes.
There is no clean fix. Text extraction will produce the same nonsense, because the information genuinely is not in the file. Rendering the page as an image and reading it, or retyping, are the practical options.
Telling them apart
Three symptoms, three causes.
| Symptom | Cause | Fixable? |
|---|---|---|
| Nothing highlights at all | No text layer - it is a scan | No |
| Whole page highlights as one block | No text layer | No |
| Text is clearly text but will not select | Permission flag | Yes |
| Text selects but pastes as nonsense | Broken character mapping | No |
| Some pages work and others do not | Mixed document | Partly |
| Selection is offset from the words | OCR text layer, approximate positions | No, but usable |
Diagnosing in under a minute
- Try to select one word
Nothing highlights: no text layer. Stop here.
- Zoom to 400 per cent
Sharp edges mean real text, so the problem is a flag rather than a scan. Blocky edges confirm a scan.
- Run text extraction
It reports how many pages produced no text, which identifies a mixed document immediately.
- If it selects but pastes badly
The character mapping is broken. Nothing in the file will fix it; render the page and read it instead.
Tools for the diagnosis
The mixed document
A document with a typeset body and a scanned appendix produces the most confusing version of this: copying works on some pages and not others, with no visible difference between them.
Text extraction resolves it in one step, because the result panel reports the number of pages that returned nothing. That number, compared against the page count, tells you exactly what kind of document you have.
Before concluding a tool is broken
- Try to select a word on the specific page you care about
- Zoom in to distinguish a scan from a flagged document
- Run text extraction and read the empty-page count
- For a flag, rewrite the file - and consider whether you should
- For a scan, plan around it; it is what the file is
Questions and answers
Why can I see text but not select it?
Almost always because the document is a scan and the text is part of an image. Occasionally because a permission flag is being honoured by your reader, which you can tell apart by zooming in - real text stays sharp.
How do I copy text from a scanned PDF?
You cannot, because there is no text. Optical character recognition would produce a machine reading of the image, with an error rate. This site does not offer it.
Why does copied text come out as gibberish?
The embedded font uses a non-standard encoding and the file lacks a correct mapping back to Unicode. The reader draws the right shapes; the file does not record which characters they are. There is no fix within the document.
Is it legal to remove a copy restriction?
For your own document, unremarkable. For material restricted deliberately by someone else so that you would not reuse it, that is a different question and one for you rather than for a tool.
