Why do journals insist that all fonts are embedded in the PDF?

Because a PDF that references a font instead of embedding it renders correctly only on a machine that already has that font — and the reviewer or typesetter almost certainly does not have your exact maths or specialist fonts. Missing-font substitution silently swaps glyphs, which is catastrophic for equations, special characters, and non-Latin scripts. Embedding bundles the actual font data into the file so it displays identically everywhere. Most submission systems check this automatically and reject files with non-embedded fonts, so verify in your PDF reader’s font properties that every font shows as embedded before you submit.

How do I anonymise a PDF properly for double-blind review?

Removing your name from the title page is not enough — author identity hides in several places. Strip the document metadata (PDF properties carry author name and the software/account that created it), remove acknowledgements and funding lines that identify you, neutralise self-citations that reveal authorship ("as we showed in [your prior paper]"), and check file properties and any tracked-changes residue. Comments and annotation authorship can also leak names. After preparing the anonymised file, open its properties and confirm the author field is blank, then read it as a stranger would to catch identity tells in the text itself.

What is PDF/A and when do I need it for a thesis?

PDF/A is the ISO-standardised archival profile of PDF, designed so a document stays openable for decades — it requires all fonts embedded, forbids external dependencies and certain encryption, and is self-contained. Many universities require the final thesis deposited to the institutional repository to be PDF/A precisely so future readers are not defeated by format drift. Journals usually want ordinary PDF for submission, but the archived version of record may be PDF/A. Check your graduate school’s deposit specification; if it asks for PDF/A, export to that profile and validate that the file actually conforms rather than assuming it does.

My PDF exceeds the portal’s size limit — how do I shrink it without ruining quality?

Compress the images, not the text. In an academic PDF the text and vector figures are tiny; the bulk is almost always raster images and embedded photos. Re-compress those to a sensible resolution (300 DPI is plenty for print, 150 for screen-only) rather than flattening or downscaling the whole document, which would blur your figures. Avoid "print to image" approaches that rasterise selectable text — they bloat the file and break searchability. If figures are still too large, supply high-resolution versions as separate supplementary files where the journal allows.

Why does the submission system say my PDF is "image-only" or not searchable?

Because the text is not real text — it is a picture of text, usually from scanning a printed page or exporting in a way that rasterised the content. Submission systems and plagiarism checkers need a selectable text layer. The fix depends on the source: regenerate the PDF directly from your writing software (LaTeX, Word) so the text is native, or, if you only have a scan, run OCR to add a text layer. Confirm by trying to select a sentence; if it highlights as text, you are fine, and if it selects as a block image, it still needs OCR or regeneration.

Is it safe to prepare an unpublished manuscript with an online PDF tool?

Unpublished research is sensitive — priority and confidentiality matter — so where the file is processed is the key question. Server-side tools upload your manuscript to a third party where it may be cached or logged; client-side (in-browser) tools compress, anonymise, and re-export locally so the file never leaves your computer. ScoutMyTool’s PDF tools run client-side. Prefer client-side processing for anything pre-publication, or use offline desktop software, and keep your own backups of the source files.

6 min read

Best PDF practices for academic submissions (PhD + journals)

By ScoutMyTool Editorial Team · Last updated: 2026-05-21

My first journal submission bounced within an hour — not on the science, but because a maths font was not embedded and the portal’s automated check caught it. I had spent months on the research and tripped over a PDF setting I did not know existed. Since then I treat the submission PDF as its own deliverable with its own checklist, because reviewers never see the file you think you sent — they see how it renders on their machine. This guide collects the PDF practices that get a paper or thesis past the automated and human gatekeepers first time: embedded fonts, size discipline, proper anonymisation, PDF/A for the archive, and the quick checks that catch desk rejections.

The submission checklist

Requirement	Why it matters	How to check
Fonts embedded	Symbols/maths render the same for reviewers	Document Properties → Fonts: all "Embedded"
Within size limit	Portals reject oversized uploads	Compress images, not text; recheck file size
Anonymised (blind review)	Author identity must not leak	Strip metadata; remove author names/acknowledgements
PDF/A for thesis archive	Long-term openability for the repository	Export as PDF/A; validate it conforms
Selectable text	Searchability, accessibility, plagiarism scan	Text selects (not a scanned image)
Correct page size	A4 vs Letter mismatches fail templates	Match the journal/grad-school template

Step by step — prepare a submission-ready PDF

Generate from source, not from a scan. Export the PDF directly from LaTeX or Word so the text is native and selectable. A regenerated PDF avoids the image-only problem that fails searchability and plagiarism checks.
Embed all fonts. Configure your export to embed fonts (LaTeX with the right driver, or Word/Acrobat export settings), then verify in Document Properties → Fonts that every font reads "Embedded" — especially maths and special characters.
Match the template’s page size and layout. Use the journal or graduate-school template, and confirm the output is the right paper size (A4 vs Letter) and margins. Mismatches fail automated template checks.
Anonymise for blind review. Strip metadata, remove author names, acknowledgements, and funding lines, neutralise identifying self-citations, and confirm the properties author field is blank. Read it once as a stranger.
Hit the size limit by compressing images. If the file is too large, re-compress raster images to a sensible DPI rather than flattening the whole document, preserving selectable text and vector figures.
Export PDF/A for the archived thesis. Where the repository requires it, export the final version of record as PDF/A and validate that it conforms, so the deposited file stays openable for decades.

The desk-rejection traps worth a second look

Most format desk-rejections come from a short list of avoidable issues: non-embedded fonts (the maths breaks on the reviewer’s screen), an image-only PDF that defeats the searchability and similarity checks, an oversized upload, a page size that does not match the template, and — for blind review — author identity that survived in metadata or self-citations. None of these touch your argument; all of them can sink a submission before a human reads a word. Building a two-minute final check around this list — fonts embedded, text selectable, size within limit, identity gone, page size correct — turns the PDF from a source of avoidable delay into a non-event, which is exactly what you want when the science is the hard part.

FAQ

Why do journals insist that all fonts are embedded in the PDF?: Because a PDF that references a font instead of embedding it renders correctly only on a machine that already has that font — and the reviewer or typesetter almost certainly does not have your exact maths or specialist fonts. Missing-font substitution silently swaps glyphs, which is catastrophic for equations, special characters, and non-Latin scripts. Embedding bundles the actual font data into the file so it displays identically everywhere. Most submission systems check this automatically and reject files with non-embedded fonts, so verify in your PDF reader’s font properties that every font shows as embedded before you submit.
How do I anonymise a PDF properly for double-blind review?: Removing your name from the title page is not enough — author identity hides in several places. Strip the document metadata (PDF properties carry author name and the software/account that created it), remove acknowledgements and funding lines that identify you, neutralise self-citations that reveal authorship ("as we showed in [your prior paper]"), and check file properties and any tracked-changes residue. Comments and annotation authorship can also leak names. After preparing the anonymised file, open its properties and confirm the author field is blank, then read it as a stranger would to catch identity tells in the text itself.
What is PDF/A and when do I need it for a thesis?: PDF/A is the ISO-standardised archival profile of PDF, designed so a document stays openable for decades — it requires all fonts embedded, forbids external dependencies and certain encryption, and is self-contained. Many universities require the final thesis deposited to the institutional repository to be PDF/A precisely so future readers are not defeated by format drift. Journals usually want ordinary PDF for submission, but the archived version of record may be PDF/A. Check your graduate school’s deposit specification; if it asks for PDF/A, export to that profile and validate that the file actually conforms rather than assuming it does.
My PDF exceeds the portal’s size limit — how do I shrink it without ruining quality?: Compress the images, not the text. In an academic PDF the text and vector figures are tiny; the bulk is almost always raster images and embedded photos. Re-compress those to a sensible resolution (300 DPI is plenty for print, 150 for screen-only) rather than flattening or downscaling the whole document, which would blur your figures. Avoid "print to image" approaches that rasterise selectable text — they bloat the file and break searchability. If figures are still too large, supply high-resolution versions as separate supplementary files where the journal allows.
Why does the submission system say my PDF is "image-only" or not searchable?: Because the text is not real text — it is a picture of text, usually from scanning a printed page or exporting in a way that rasterised the content. Submission systems and plagiarism checkers need a selectable text layer. The fix depends on the source: regenerate the PDF directly from your writing software (LaTeX, Word) so the text is native, or, if you only have a scan, run OCR to add a text layer. Confirm by trying to select a sentence; if it highlights as text, you are fine, and if it selects as a block image, it still needs OCR or regeneration.
Is it safe to prepare an unpublished manuscript with an online PDF tool?: Unpublished research is sensitive — priority and confidentiality matter — so where the file is processed is the key question. Server-side tools upload your manuscript to a third party where it may be cached or logged; client-side (in-browser) tools compress, anonymise, and re-export locally so the file never leaves your computer. ScoutMyTool’s PDF tools run client-side. Prefer client-side processing for anything pre-publication, or use offline desktop software, and keep your own backups of the source files.

Citations

Compress your manuscript to the size limit

ScoutMyTool Compress PDF shrinks image-heavy manuscripts in your browser — selectable text and vector figures preserved, nothing uploaded. Get under the portal limit without sacrificing figure quality.

Open Compress-PDF tool →