Students mostly need three things from a file tool: turn a pile of phone photos into one submission-ready PDF, get copyable text out of a scanned chapter, and squeeze a file under an upload limit. All three can be done in the browser, without installing anything and without uploading your work.
That last point matters more than it sounds. Coursework, transcripts, medical notes and draft theses are exactly the kind of files people are uneasy about handing to a random website, and campus networks are often slow enough that a 40 MB upload is its own problem. A browser tool reads the file from your disk, does the work on your own processor and hands the result straight back, so nothing is transmitted and there is no upload wait. No account is involved either, which is one less thing to remember at 2am before a deadline.
Turn photos of pages into one PDF
The most common student file is not a document at all, it is a set of phone photos of a handout, a library book or a whiteboard. Submitting eight separate JPEG files is usually not accepted, so they need to become one PDF in the right order.
Add every photo at once, drag the cards until the order matches the pages, then choose A4 or Letter, the orientation, the margin and how each image should fit the page. JPG and PNG files are embedded as they are, and WebP, GIF or AVIF images are converted first. Each image becomes one page, and the file never leaves your device.
There is no image count limit. The practical ceiling is your own memory, since every photo is held locally while the PDF is assembled.
Image to PDF
Drop in the photos, set the page order, size and margins, then download one PDF. Nothing is uploaded.
Get copyable text out of a scanned chapter
A scanned reading list is a picture of a page, so selecting the text does nothing. You have two routes depending on what you already have.
If the pages are photos or screenshots
Use Image to Text. Add one image or several, pick the language of the text, and run the recognition. English is built in, and eleven more languages including Chinese, Japanese, Korean, German, French, Spanish, Russian and Arabic can be selected from the language menu. The first run downloads about 6.5 MB of recognition engine and language data, and your browser caches it, so later runs start almost instantly.
If the pages are already in a PDF
Use PDF to Text, which handles both cases. Pages that carry a real text layer are lifted out directly, which is instant and perfectly accurate. Pages that are just a picture are rendered and recognised with OCR instead. The summary next to the result tells you how many pages came from each route, so you know which part of the output deserves a closer proofread.
PDF to Text
Reads the text layer instantly and runs OCR on scanned pages, all in the browser.
Fit a file under an upload limit
Portals and email gateways cap attachments, and a scan exported at full quality can easily exceed them. Two tools cover almost every case.
Compress PDF rewrites the internal streams of a PDF using qpdf compiled to WebAssembly. The Standard setting repacks them at compression level 6, and Maximum uses level 9 and also linearizes the file so a reader can show the first page before the download finishes. The tool is designed for files under 50 MB.
Image Compressor handles photos and screenshots with a quality slider and a live before and after size, and it can re-compress a PNG losslessly with OxiPNG.
One caveat worth stating plainly: neither tool takes a target size. You cannot type 100 KB and get exactly that. You lower the setting until the preview shows a size you can live with, which is the honest version of what these tools do.
What these tools cannot do
- No handwriting recognition. OCR is trained on printed text, so handwritten lecture notes will produce mistakes. Treat the output as a draft.
- No editing of existing PDF text. You cannot reword a paragraph inside a PDF. You can add page numbers, a watermark or a signature on top of it, and you can reorder or merge pages.
- No unlimited compression. Push scanned text too far and it stops being readable. There is a floor, and it arrives before the file size stops falling.