A school office handles the same document twice: once as a stack of paper and once as a file that has to reach a portal, a board or a family. The chain is read the scan, split it into one file per student, merge each pack in order, number the pages, compress to the portal limit and encrypt the copy that leaves the office. Everything runs in the browser.
The split comes before the merge because a scanner produces one long file, not one file per applicant. Page numbers come after the merge for the same reason they do in a book: numbering the sections separately gives every pack a page one. Compression comes before encryption, because no other tool here can open a protected file, and encryption comes last because nothing downstream can read it. That order also keeps the unencrypted master available for the next request.
1. Read the scan before you split it
PDF to Text reads the existing text layer first and runs OCR only on pages that have fewer than 20 characters after whitespace is removed, which is how a scanned page is detected. Those pages are rendered at twice their size for recognition. The tool supports 12 languages, marks each page in the output and exports a .txt file.
Two limits matter. The output is a text file, not a searchable PDF: this tool does not add an OCR layer to the original pages, so the PDF itself stays an image scan. And the recogniser is reliable on printed forms, not on handwriting, signatures or a form photographed at an angle. Use the text to check that the scan is complete and in order before you cut it.
PDF to Text
Read a scanned pack and export a text copy with page markers.
2. Split one scan into one file per student
Split PDF has three modes: page ranges such as 1-3, 5, 8-10; every page as its own file; and fixed chunks of N pages. Fixed chunks is the mode for a batch of application forms where every student has the same page count. Scan the stack in list order, set the chunk size to that count, and the tool produces one file per block. One run can output up to 200 files, and page numbers inside each file start at one again.
The setting is taken literally. One missing or duplicated sheet shifts every file after it, so check the first and last output before you delete the master scan. If the intake is longer than 200 files, split the stack into two scans. The tool handles one PDF at a time and cannot queue several files.
Split PDF
Fixed chunks turn a stacked scan into one file per application.
3. Merge each pack in reading order
Most packs are more than the application form: a referee letter, a transcript, a proof of address and a signed consent. Merge PDF joins them in the order you set, needs at least two files, copies pages as they are and keeps links intact. It does not compress, so a pack of full page scans stays the size of its parts.
An encrypted PDF has to be unlocked first, and a merge is a new copy, so a corrected page means rebuilding from the sources. There is no file count or size cap in the code, but a very large pack is limited by the memory of the browser tab.
Merge PDF
Join the form, references and supporting pages into one pack.
4. Number every page
Add Page Numbers offers four formats, a starting value, six positions and an option to skip the first page. Font size runs from 8 to 24 pt in four typefaces, and the numbers are drawn as text, so a reviewer can select and search them. Use 1 / N or Page 1 of N when someone will refer to a specific page, and skip the first page when it is a cover sheet.
There is no prefix field, so a student name or an application number cannot travel inside the number. If each pack needs an identifier, put it on a cover page in the source document before merging. The tool handles one PDF at a time, which is the slow part of a large intake.
Add Page Numbers
Continuous numbering with an optional skipped cover.
5. Compress to the portal limit
Compress PDF is designed for files under 50 MB. It rewrites the internal structure with qpdf, packs objects and rebuilds the cross-reference table, so text and links survive, but the scans inside keep their original size. A pack of 300 dpi colour scans barely moves.
The real saving happens at the scanner: scan in greyscale at 200 or 300 dpi when the portal accepts it, and keep each pack to the pages it needs. There is no target size field, so run the tool and read the number. If it is still too heavy, send two packs rather than degrading every page, and keep one uncompressed master for records.
Compress PDF
Structural compression under a 50 MB cap, without touching the scans.
6. Encrypt the copy that leaves the office
Protect PDF sets AES-256 with an open password and optional permissions for printing, copying and editing, entirely in the browser. The minimum password length is four characters, and the file is capped at 50 MB. Send the password by a different channel than the file, and keep the unencrypted master.
Three things it is not. It is not redaction: a name printed on a page is still readable once the file is open. It is not a signature: there is no certificate, no proof of who produced the pack and no timestamp. And it is not recoverable: a lost password loses the file for everyone. Encryption is one control in a records policy, not a compliance certificate on its own.
Protect PDF
AES-256 with print and copy permissions, on your own device.
What this chain cannot do
- No form filling or field extraction. OCR returns text, not a spreadsheet of names and grades.
- No searchable PDF. PDF to Text exports a .txt file and leaves the original scan unchanged.
- No handwriting or signature recognition. Printed forms only.
- No redaction. PDF Metadata clears six document fields; it cannot black out a name or a grade on a page.
- No batch numbering or batch encryption. One PDF at a time.
- No student database, mail merge or document templates. Every pack is built from the files you supply.
- No ZIP packaging. Each output is downloaded on its own.
- No HEIC or TIFF intake. Convert those before they enter the chain.
- No certificate signature. Sign PDF places a drawn or uploaded image, not a cryptographic signature.