A PDF stores glyphs, the shapes to draw, not text. What you get when you copy comes from a separate mapping inside each font that says which letter each glyph stands for. When that mapping is missing or wrong, the page looks perfect and the pasted text is boxes, symbols or the wrong letters. This tool finds out whether that is the case and which font is responsible.
It is not a PDF-to-text converter: it does not hand you the text. It tells you whether copying works, page by page, why it does not, and what to do about it. The PDF is analysed in your browser and is not uploaded.
Loading the tool…
Overview
About PDF Copy-Paste Text Health Checker
Two signals. First, the fonts: for each font the tool reads the dictionary and, when present, the /ToUnicode CMap (ISO 32000-1, 9.10.2), counts its mappings and checks where they lead. It also recognises font kinds that copy badly: Type 3 fonts, which draw glyphs from PDF instructions and have nothing to map; composite (CID) fonts with Identity encoding and no ToUnicode map, whose codes are glyph numbers; encodings that name glyphs generically (g12, glyph034) without a ToUnicode map; and names of legacy Bangla fonts such as SutonnyMJ, which copy as Latin-looking symbols.
Second, the text itself: the tool extracts each page with the pdf.js reader and scores it.
The readable-text score. For a page, the share of non-space characters that are ordinary letters, digits, punctuation or symbols, after removing the characters that are never ordinary text: the replacement character U+FFFD, private-use code points, control characters, unassigned code points, "(cid:123)" placeholders and the typical pattern of UTF-8 text read as Latin-1 (such as "é"). A page below 97 gets a warning, below 80 it is marked unreadable, and text whose words run together or whose letters are spaced apart is flagged too. The score is a heuristic: it cannot see letters that are valid but wrong, which only the font diagnosis and the preview can show.
What you see. A verdict, the reasons (font by font and page by page), a table of fonts with their copy risk, and for each page a preview of what copying gives with the list of unreadable characters, such as U+E0A1 ×12. A note appears when copied text contains ligature characters like "fi", which can break a search for "fi". Pages without text are named, since a scan has nothing to copy until it is run through OCR.
Related. The PDF Font Checker lists fonts, embedding and subsets; this tool uses that information to answer the copy question. For legacy Bangla text, the Bijoy PDF to Unicode tool converts it.
Read more. The guide Why text copied from a PDF turns into symbols explains ToUnicode maps and what to export differently.
Compatibility
Supported Formats
PDF files up to 80 MB; the first 150 pages are checked. Password-protected PDFs are refused with a message. Output is on screen and a JSON report; no text is exported.
Guide
How to Use PDF Copy-Paste Text Health Checker
- Drop a PDF onto the box. Pages are read and scored in your browser.
- Read the verdict and the readable-text score, then the reasons: fonts and pages named.
- Open the page previews to see what copying gives and which characters are unreadable.
- Fix the cause at the source (export with fonts embedded and a ToUnicode map, avoid outlines or Type 3 fonts, or run OCR for scans), then check the new file.
Highlights
Features
- Font diagnosis ToUnicode maps, Identity-encoded CID fonts, Type 3 fonts, generic glyph names and legacy Bangla fonts.
- Readable-text score A defined score per page: share of ordinary characters after removing replacement, private-use and control characters.
- Copy preview What copying gives on each page, with the unreadable characters listed by code point.
- Page by page Names the pages and the font behind each problem, and pages without text.
- No text export A diagnosis, not a converter: nothing is uploaded and no text file is made.
Applications
Use Cases
- Pasted text from a PDF is symbols, boxes or the wrong letters
- Checking a PDF you publish so readers and screen readers get proper text
- Finding the one font that breaks copying in a long document
- Telling a scan from a text PDF before searching it
- Spotting legacy-encoded Bangla text before converting it
Good to Know
Limitations
Text is extracted with the pdf.js reader. Other viewers (Acrobat, pdfium, Preview) differ in how they use mappings, so a page this tool rates healthy can still copy badly somewhere, and the reverse. The score cannot detect valid-but-wrong letters, and the font diagnosis flags risk patterns rather than proving a failure.
It does not repair the file or change fonts, and text converted to outlines or drawn as pictures has nothing to copy at all, which the tool reports as "no selectable text".
This tool runs entirely in your browser — your file is never uploaded to our servers. See our Privacy Policy for full details.
Support
Frequently Asked Questions
Why does copying from my PDF give symbols or boxes?
The font has no usable mapping from its glyph codes to Unicode (no ToUnicode map, or a wrong one), so the viewer cannot tell which letter each glyph is. The checker shows which font and which pages.
Is my PDF uploaded?
No. It is read in your browser and not sent anywhere.
What is a ToUnicode map?
A table inside a font dictionary that maps glyph codes to Unicode text. Copy, search and screen readers use it. Without it, a viewer has to guess from the encoding or the font file.
This is not a PDF-to-text converter?
Correct. It diagnoses copying and shows a short preview per page, but does not give you the text. Use a converter for that once the cause is fixed.
The score is 100 but the copied text is wrong. Why?
The score only counts characters that are not ordinary text. A font that maps letters onto other real letters produces valid but wrong text. The font table and preview are there to catch that, and legacy Bangla fonts are named explicitly.
The PDF is a scan. What then?
A scan holds pictures, not text, so the checker reports no selectable text. Run OCR to add a text layer, then check that layer with the PDF OCR Layer Alignment Checker.
Keep Going
Related Tools
- PDF to Text Extract the real text layer from a PDF as a plain .txt file — free and fast.
- PDF Text Density Checker Check whether a PDF's pages have real extractable text or are likely scanned images — free and instant.
- Bijoy / SutonnyMJ PDF to Unicode Text Read the text of a PDF made with a legacy Bangla font (Bijoy, SutonnyMJ) and convert it to Unicode Bangla. The raw text and the result sit side by side. Runs in your browser.
- PDF Font Checker List the fonts in a PDF: type, embedded or not, subset, encoding, and the pages that use them. Warns about fonts that are missing from the file. Export a report. In your browser.
- PDF OCR Layer Alignment Checker Check whether the hidden, searchable text of a scanned PDF sits on the words in the picture: overlay view, offset and rotation detection and a page-by-page score. In your browser.