How to split a pdf every n pages for archive and storage
- Step 1Confirm your DMS per-document limit — Check whether your archive caps by page count or file size (or both). Translate a file-size cap into pages using the source's average page weight (total MB ÷ total pages), since this tool cuts by page count, not bytes.
- Step 2Open the tool and load the archive PDF — Drop the long document into PDF Split (Fixed). The page count appears once it's read — note it, you'll use it to predict the volume count.
- Step 3Set the volume size in Pages per chunk — Enter the maximum pages per volume — e.g. 100 — in the single Pages per chunk field. There are no presets; any whole number ≥ 1 works.
- Step 4Process and verify the volume count — Click Process. The result panel shows Output files = ceil(total ÷ volume size). Confirm it matches your expectation before downloading, so a mistyped volume size doesn't slip through to ingestion.
- Step 5Download the numbered volumes — Click Download. Volumes save sequentially as name.split-fixed.1.pdf, name.split-fixed.2.pdf, …. They are separate files (no zip), staggered ~200 ms apart.
- Step 6Rename to your retention scheme and ingest — The tool can't apply your classification naming, so rename the downloads to your records convention (e.g. CASE-2026-001_v01.pdf). Run PDF OCR on each volume if your archive requires searchable text, then ingest in order.
Choosing a volume size
Trade-offs between fewer large volumes and many small ones. Volume count is ceil(total pages ÷ pages-per-volume).
| Pages per volume | Best when | Volumes for a 600-page archive | Trade-off |
|---|---|---|---|
10 | Strict small-object DMS, fine-grained retrieval | 60 | Many files to track and name |
50 | Balanced indexing vs. file count | 12 | Good general default |
100 | Generous page cap, fewer objects | 6 | Larger per-volume file size |
250 | Loose cap, minimal volume sprawl | 3 | Each volume heavier to open |
Converting a file-size cap into a page count
This tool splits by pages, not bytes. Estimate pages-per-volume from the source's average page weight so each volume lands under a file-size cap.
| DMS file-size cap | Avg page weight | Safe pages per volume | Note |
|---|---|---|---|
| 50 MB | ~0.5 MB (text scan) | ~90 | Round down for headroom |
| 50 MB | ~2 MB (colour scan) | ~22 | Image-dense pages weigh more |
| 20 MB | ~1 MB | ~18 | Verify the largest volume after splitting |
| 100 MB | ~0.3 MB (born-digital) | ~300 | Page cap may bind before size cap |
Cookbook
Volume-size recipes for common archive policies. Volume count is always ceil(total ÷ pages-per-volume).
Every 100 pages for a 600-page case file
A records policy caps each ingested object at 100 pages. Set Pages per chunk to 100 for six clean volumes.
Input: case-file.pdf (600 pages) Pages per chunk: 100 Output files panel: 6 case-file.split-fixed.1.pdf pages 1-100 ... case-file.split-fixed.6.pdf pages 501-600
Every 10 pages for fine-grained retrieval
A small-object DMS wants 10-page volumes so individual records are quick to pull. 250 pages → 25 volumes.
Input: ledger.pdf (250 pages) Pages per chunk: 10 Output files panel: 25 ledger.split-fixed.1.pdf pages 1-10 ... ledger.split-fixed.25.pdf pages 241-250
File-size cap translated to pages
DMS caps objects at 50 MB. The colour-scan archive averages 2 MB/page, so ~22 pages fits under 50 MB with headroom.
Source avg: 2 MB/page Cap: 50 MB -> ~22 pages Pages per chunk: 20 (rounded down for safety) Output files panel: ceil(total / 20) volumes, each ~40 MB
OCR each volume after splitting
Split first, then make each scanned volume searchable individually — OCR runs per file.
Step 1: Split (Fixed) -> archive.split-fixed.1..N.pdf
Step 2: for each volume:
PDF OCR /pdf-tools/pdf-ocr (lang: eng)
-> searchable text layer addedRecombine volumes for export
Need the full record back for a disclosure request? Merge the volumes in order.
Tool: PDF Merge /pdf-tools/pdf-merge Add: case-file.split-fixed.1.pdf ... .6.pdf (in order) Merge -> single 600-page record
Edge cases and what actually happens
Last volume shorter than the rest
By designIf the page count isn't a multiple of the volume size, the final volume holds the remainder. A 250-page archive at 100 per volume gives 100, 100, 50. This is expected and complete — no pages are dropped or padded.
DMS caps by file size, not page count
estimate pagesThis tool splits strictly by page count. To meet a megabyte cap, estimate pages-per-volume from the source's average page weight (total MB ÷ total pages) and round down for headroom. Verify the largest output volume against the cap, since image-dense sections weigh more than average.
Original Title/Author metadata not in each volume
ExpectedEach volume is a new PDF assembled from copied pages, so document-level metadata and bookmarks from the original don't propagate. Re-apply archival metadata in your DMS, or use the Metadata Scrubber if you need a clean, known metadata state per volume.
Existing OCR text layer survives the split
PreservedBecause pages are copied rather than re-rendered, an invisible OCR text layer already on the pages is carried into the volumes intact. (Contrast with lossy compression, which would discard it.) If pages have no text layer yet, OCR each volume afterward with PDF OCR.
Free tier — file over 2 MB
blocked (upgrade prompt)Archive PDFs are usually well over the Free 2 MB ceiling, so they're blocked at upload on Free with an upgrade prompt. Pro raises this to 50 MB, Pro+Media to 500 MB, Developer to 2 GB. The check runs before any split.
Free tier — more than 50 pages
blocked (upgrade prompt)Free caps PDFs at 50 pages, so most archive documents trip the page block with an upgrade prompt. Pro lifts it to 500 pages, Pro+Media to 2,000, Developer to 10,000 — relevant when ingesting large scanned records.
Volumes are separate files, not a zip bundle
ExpectedEach volume downloads individually (staggered ~200 ms) with no zip packaging. Your browser may prompt to allow multiple downloads — approve it. Organise the saved volumes into your archive folder structure afterward.
Custom retention naming not applied automatically
rename afterFiles come out as name.split-fixed.1.pdf etc.; the tool can't apply your classification or retention naming convention. Rename the downloads in your file manager (or a batch-rename utility) to match your records scheme before ingestion.
Password-protected archive PDF
Often supportedLoaded with ignoreEncryption: true, so many protected archives split fine. If pdf-lib can't parse the encryption, remove the password first with PDF Remove Password, then split into volumes.
You need cut points at record boundaries
wrong toolFixed split cuts at equal page intervals, ignoring where one record ends and the next begins. If your archive must break on logical boundaries (one case = one volume), use PDF Split by Range and specify each range explicitly.
Frequently asked questions
What if my DMS limits by file size rather than page count?
This tool splits by pages, so estimate a safe pages-per-volume from the average page weight: total MB ÷ total pages gives MB per page, then volume cap ÷ MB-per-page gives pages, rounded down for headroom. A 50 MB cap on a 2 MB/page colour scan means about 22 pages per volume. Verify the largest output volume against the cap, since image-heavy pages weigh more than average.
Will the volumes lose image quality?
No. Pages are copied with pdf-lib copyPages — there's no re-rendering or re-compression — so scanned image bytes are carried into each volume exactly. The split is lossless; any quality reduction would have to be a separate, deliberate compression step.
Does an existing OCR text layer survive splitting?
Yes. Because pages are copied verbatim, an invisible OCR text layer already present on the pages is preserved in the volumes. If your scans have no text layer yet, run each volume through PDF OCR after splitting to make them searchable.
Will metadata and bookmarks carry over to each volume?
Page content carries over, but document-level metadata (Title, Author) and bookmarks generally do not, because each volume is a new PDF built from copied pages. Re-apply archival metadata in your DMS, or normalise it with the Metadata Scrubber.
Can I run OCR on the volumes after splitting?
Yes — run each volume through PDF OCR individually to embed a searchable text layer over the scanned images. OCR is per-file, so process the volumes one at a time; the first run downloads a small language pack that's cached afterward.
How many volumes will I get?
The volume count is the total page count divided by your pages-per-volume, rounded up: ceil(total ÷ size). A 600-page archive at 100 pages per volume produces 6 volumes; the result panel shows the exact Output files count before you download.
Are the volumes bundled into a single zip?
No. Each volume downloads as its own PDF, one after another with a small stagger. There is no zip output. If your browser blocks multiple downloads, approve the prompt, then organise the files into your archive structure.
Can I rename the volumes to my records convention?
Yes, after download. The tool names files name.split-fixed.1.pdf, .2.pdf, and so on; it does not apply your classification scheme. Rename them in your file manager or a batch-rename tool — e.g. CASE-2026-001_v01.pdf — before ingestion.
Is my archive uploaded during splitting?
No. The split runs entirely in your browser and the panel reports "0 bytes uploaded". For records-management originals subject to retention or chain-of-custody requirements, the file never leaves your machine, which avoids a third-party-disclosure surface.
What if records cross my fixed page boundaries?
Fixed split cuts at equal intervals regardless of content, so a record can straddle two volumes. If your policy requires one logical record per object, use PDF Split by Range and enter the exact ranges that match each record's boundaries.
Can I recombine volumes later for a disclosure request?
Yes — PDF Merge recombines the numbered volumes into the full record. Add them in numeric order so page sequence is restored, then merge and download a single PDF for export.
How large an archive can I split per job?
Free is limited to 2 MB and 50 pages — too small for most archives. Pro allows 50 MB and 500 pages, Pro+Media 500 MB and 2,000 pages, Developer 2 GB and 10,000 pages. Limits are checked on upload, so an oversized archive is flagged before you set the volume size.
Privacy first
All PDF processing runs locally in your browser using PDF-lib and pdf.js. No file is ever uploaded — only metadata counters are saved for signed-in dashboard stats.