The sizes are not predicted from page counts. Bindery writes each candidate part and measures it, because a page with a photograph weighs a hundred times more than a page of text and no estimate survives that. The parts are named after the original with a part number, packed into a single ZIP on your device, and nothing is uploaded at any point. If the parts turn out numerous, run Compress first — Balanced often halves a scan without visible change — and split the result. Files above 200 MB open in big-file mode, which streams splits by page count; the size-based split needs the file in memory.
How it works
- Drop the PDF, open Split and choose “under a size”
- Enter 5 MB — the board shows which pages fall into each part
- Split; every part is measured as it is written and downloads in one ZIP
Questions people ask
How does Bindery know a part is really under 5 MB?
It writes the part and checks the byte count, adding pages until the next page would push it over the limit. That makes the result exact rather than estimated, at the cost of a little more work on your device — which is fine, because nothing has to travel to a server.
Can I compress the file instead of splitting it?
Often both help. Compress resamples the images and shows the real output size of each level first; if that lands under 5 MB you are done with one file. If not, split the compressed result and you will get fewer, smaller parts than by splitting the original.
How does the recipient put the parts back together?
Any merge tool will do, including Bindery’s, which copies pages unchanged and writes a real outline. Because the parts are consecutive page ranges named in order, merging them in file-name order reproduces the original document exactly. Nothing needs to be re-sent if a part goes missing; split the original again and send that one part.