2026-09-01

Grad application season opens September 1 — how to excerpt your writing sample and compile transcript PDFs, without uploading them anywhere

Grad application season has an actual opening bell this year, and it rings today: UC Riverside's graduate division states that "the Summer/Fall 2027 application opens on September 1, 2026 to new applicants, for most graduate programs" — and across the country, other portals open on roughly the same schedule, with deadlines clustering in December and early January (UCR's own fellowship-consideration deadline is January 5). Between now and then, two upload boxes cause most of the purely mechanical grief: the writing sample and the transcripts. Neither is intellectually hard — the paper is already written, the transcript is already issued — but both come with picky shape requirements that programs spell out in their own documentation and that a word processor alone doesn't produce. And application files are exactly the kind of thing that shouldn't detour through a free converter site's server: your grades, and research you haven't published yet.

What the upload boxes actually ask for

The writing sample slot is not "attach your paper." The University of Chicago's Division of the Social Sciences spells out a shape many programs share: 20 to 30 pages for PhD programs (15 to 20 for MA programs), excluding bibliography or appendices, with anything longer than 30 pages "read only at the discretion of the departmental review committee." For longer works like a thesis, the instruction is to "provide an excerpt shorter than 30 pages" plus "a one-page coversheet contextualizing the excerpt within the larger body of work" — and if you're submitting more than one piece, "multiple writing samples must be combined into a single pdf before upload." The transcript slot has its own geometry. CU Boulder's Graduate School puts it in one sentence: "Make one compiled transcript PDF for each institution you attended, even if it includes multiple pages, translations, a legend, degree certificate, or diploma" — with a hard cap of 1.5 MB per compiled PDF, and a blunt warning that "secure PDFs and non-legible scans are not acceptable." Purdue's graduate school says the same thing from the other side: "Please do not upload a document that is password-protected or that contains macros," keep the scan fully legible on letter-size paper, and scan at 200 dpi or less. Add it up and the pattern is consistent: per slot, one PDF, unencrypted, readable, in the right order, under the size cap.

Excerpting a writing sample without re-editing the paper

If your sample lives in a word processor, the cleanest excerpt comes from the source: cut the document down, let the pages renumber themselves, export a fresh PDF. But a lot of real writing samples aren't editable anymore — the seminar paper PDF a professor graded, the published article downloaded from the journal, the thesis chapter typeset in LaTeX two laptops ago. For those, NearPDF's delete tool does the excerpting on the file itself: every page appears as a thumbnail, clicking one opens a full-size preview so you can confirm exactly where the excerpt should start and end, and each thumbnail has its own delete button. Pages you keep stay in their original order. Then write the one-page coversheet UChicago asks for in any word processor, export it as its own small PDF, and use the merge tool to put coversheet and excerpt into the single file the upload box wants — the merge grid shows both files' pages together, so you can confirm the coversheet sits in front before you commit. One honesty note on page numbers: NearPDF's page-number tool stamps a fresh 1, 2, 3 at the bottom center of each page, but it doesn't detect or remove numbers the paper already has — an already-numbered excerpt would end up with both sets. A typeset paper almost always is already numbered, so leave it alone (the coversheet explains where the excerpt came from anyway) and save the stamp for pages that genuinely have no numbers at all.

One compiled transcript PDF per institution

The transcript slot is a merge job by definition once your paper trail involves more than one piece of paper: CU Boulder's "one compiled PDF per institution" explicitly includes translations, the grading legend from the back of the transcript, and the degree certificate in the same file. So the assembly is: scan whatever exists only on paper with a scanning app, collect what's already digital, drop the pieces for one institution into NearPDF's merge workspace, drag them into a sensible order — transcript first, legend behind it, translation and degree certificate after — and rotate any sideways capture with the per-page rotate control before merging (rotation in NearPDF changes a page property rather than re-rendering the page as an image, so a text transcript stays selectable and legible; the rotation post covers why fixing it in the file beats hoping the reviewer's viewer does). Repeat per institution. The result is exactly the shape the checklist describes, built without your grades ever leaving your device — and if the same institution's scans are also headed for a job application later, the job-application post covers that variant of the same merge.

The secured e-transcript trap

There's one failure mode worth knowing before deadline week, because it looks like a bug and isn't. Official e-transcripts from vendors like Parchment often arrive as encrypted PDFs, and a registrar help-desk knowledge-base article at Bowling Green State University documents the consequence directly: "due to the encryption that's on the PDF file, you are not able to upload the file" to another university's application — and the encryption isn't removable on request, because per the vendor the files "have to be sent securely by law." That's the same lock NearPDF hits: an encrypted PDF is skipped from a merge automatically, with a per-file notice naming it, because NearPDF can't open — let alone decrypt — password-protected files. The workaround is the one the registrar KB itself gives: print the transcript and rescan the pages as a new, unencrypted PDF (with a scanning app, not NearPDF), then compile that rescan with the rest of the institution's documents. And remember what the unofficial-upload slot is actually for — CU Boulder wants the uploaded copy for "review and decision purposes," while the sealed official transcript follows separately after admission, so a clean rescan of your own copy is precisely what this slot expects.

The honest limits

NearPDF won't turn photos of a paper transcript into a PDF — its file intake accepts only PDFs, so the scanning app comes first. It has no compression, which matters for CU Boulder's 1.5 MB-per-compilation cap: deleting pages shrinks a file only by the pages removed, so if a compiled transcript is still too heavy, the fix is at capture time — rescan at lower resolution (Purdue's own tip is 200 dpi or less) or in black-and-white. It has no OCR and can't sharpen an illegible scan, and "non-legible scans are not acceptable" is written into the same checklist as the size cap. It cannot remove a password from a secured e-transcript — it can only tell you, per file, that one was skipped. And it knows nothing about your programs' actual rules: page caps, coversheet expectations, and how each portal wants multiple institutions handled all live on each program's own admissions page, which outranks everything here. What NearPDF does is the mechanical middle: a 42-page thesis, a sideways degree-certificate scan, and a loose translation in — a page-capped excerpt with its coversheet, and one compiled transcript PDF per institution out. Built in your browser, on your device, with your grades and your unpublished research uploaded to nothing except the application portal itself.