Split a large PDF into several files
A large file stops you in one of two places: a portal refuses it for being over the limit, or the document is long and you only want part of it. Both are answered by cutting rather than compressing, because cutting copies the pages exactly as they are and touches nothing inside them. Drop your file into the tool below and choose what comes out, then read the sections underneath to see which kind of cut your case needs, and what a part keeps and drops. Everything runs inside your browser on your own device, so the file is never uploaded, and there is no account and no daily limit.
Drop your files here
Files are processed on your device and never uploaded
Runs in your browser. Your file never leaves your device.
You can process several files at once, up to 3.
Large means two different things, with two different cures
Separate the two before you cut. The first file is large in megabytes: thirty scanned pages weighing 20 MB against a portal that wants under 2 MB. What is in your way is a number in megabytes, and the right cut is the one that counts bytes and keeps every part under the ceiling. The second file is large in pages: a four hundred page book that is only 3 MB, where what is in your way is that you want chapter three on its own, or you have to hand each chapter over separately. There the number is a page count, and the right cut is the one that counts pages. The tool on this page cuts by pages and answers the second case; the split by size page linked below answers the first. Plenty of files are both, so cut by pages first and compress whichever part is still heavy.
Several files in one run, not one run per file
This is the option people mean when they search for splitting a PDF into several files. Choose Several files under what comes out, then tap the pages where each new file begins, or type their numbers into the range field separated by commas. Page 1 always begins the first file, so you never name it, and each part runs to the page before the next part starts. A 96-page file cut at 25, 49 and 71 comes out as four files covering 1-24, 25-48, 49-70 and 71-96. Before you press the button the list of files appears under the thumbnails, drawn from the same function the engine cuts with, so what the page promises beforehand and what the archive holds afterwards cannot drift apart. The parts arrive in one ZIP, and each carries its page numbers in its name, like part-2-pages-25-48.pdf, because you are most likely about to upload them one at a time into a form that asks which pages are in which.
When the limit is a number in megabytes
If the problem is that a portal refuses the file for its size, cutting by pages is guesswork: nobody knows how many pages weigh 2 MB. Open the split by size page linked below and pick the ceiling from the five offered (1, 2, 5, 10 and 25 MB). Each part is filled with the most consecutive pages that keep it under the ceiling before the next one starts, so you get the fewest parts possible and no part comes out half empty. And the number is not estimated: every candidate part is actually written and its bytes counted, because a PDF's size is not the sum of its pages, an embedded font or a repeated image being stored once however often it is used. If one page is bigger than the ceiling on its own, cutting has no answer for it, since the smallest possible part is a single page; it goes out as its own part, marked, with its size shown, so you do not upload it and get rejected without knowing why.
Writing the page range the way you actually type it
The range field takes numbers, commas and dashes: 1-3,5,8 means pages 1 through 3, then 5, then 8, and an empty field means the whole file. Four details get asked about constantly and all four are tested. An open dash works: 50- means page 50 to the end of the document however long it is, which is the quickest way to lop the tail off a long file. A range that runs past the last page stops at the last page, so 90-200 in a 96-page file gives you 90-96 rather than nothing. A number that falls entirely outside the file, like 200 in a 96-page file, is refused with a message saying the range is invalid, instead of producing a file with no pages in it. And typing on an Arabic keyboard is taken as it comes: Arabic-Indic digits work, the Arabic comma works, the long dash that autocorrect substitutes for a plain hyphen works, and spaces make no difference.
The order of the pages follows the order you typed
If you choose One file under what comes out, every page you named arrives in a single file called selected-pages.pdf, and their order in it is the order you typed rather than their order in the original. Type 5,1 and page five comes first with page one behind it. That is the quickest way to bring a page forward without opening the organize tool, and it is also the easiest mistake to make if you type the numbers as you remember them rather than as you want them. One difference matters here: tapping thumbnails writes the range back sorted ascending and de-duplicated, because a tap has no order of its own, so if you want an order that differs from the document's, type it into the field by hand after you finish tapping.
Every page as its own file, and when not to choose it
The third option, Every page, makes each page a file of its own, all of them in one ZIP, each named for its page number like page-7.pdf. A ZIP opens with one tap in the Files app on iPhone and Android and on any computer with no extra software. This is the right option when every page really is a separate document: a batch of certificates, a day-by-day statement, or invoices collected into one file. It is the worst possible option for a book: four hundred pages means four hundred files in one archive and no way to tell which one is chapter three. For books, reports and chapters choose Several files and tap the chapter openings instead, and you get eight files named for their pages rather than four hundred named for nothing.
What a part keeps, and what it leaves behind
Page content travels exactly as it is: text, fonts, images, orientation and resolution, with no recompression and nothing redrawn, so a page in a part is the page that was in the original. What belongs to the document as a whole rather than to any one page does not travel, and that is measured rather than reasoned about: a test file with a title, an author and one filled form field came out, in its first part, with no title, no author and zero fields. The outline and bookmarks belong to the document in the same way. What to do about it: if you have a filled form, run it through the flatten tool before cutting, so the answers become permanent page content and travel with the page. If the title and author matter in the parts, write them afterwards with the file details tool.
Splitting does not compress, and the parts add up to more
This surprises people and it follows from how a PDF is built: cutting does not shrink the document, it distributes it. More than that, the sizes of the parts add up to more than the original, because an embedded font or a repeated logo that was stored once in one file is now stored once in every part. In a measurement here, a six-page file cut in half produced two halves whose combined size was about a third larger than the file they came from. What matters to you is that each part on its own is smaller than the original, and that is the number the portal is asking about. If a part is still over the limit after the cut, compress it at the light level first, and leave the strong level for scanned pages, since it redraws each page as an image and can hand a scanned contract back with smeared letters.
A large file on a small device, and a password-protected one
There is no ceiling imposed from our side, neither on file size nor on page count, because the work runs on your device rather than on a server of ours, and the only limit is the device's own memory. Cutting is among the lightest things these tools do: pages are copied as they are and nothing is redrawn, unlike strong compression or OCR, which render every page again. Thumbnails are drawn gradually, nearest to your screen first, so a 300-page file does not paint 300 pictures before you can see the first one. A password-protected file is a separate case: it refuses to open and you see a message that it is protected, which is encryption rather than a fault. If you know the password, run it through the unlock tool first and cut the result, and the same goes for a payslip that arrives from an HR system locked with an ID number or a date of birth.
Related pages
Three steps, that's it
Drop the large PDF into the tray
Choose Several files, then tap the pages where each file should begin
Check the list of files under the thumbnails, press split and download
Frequently asked questions
How do I split a large PDF into several files?
Drop the file into the tray, choose Several files under what comes out, then tap the pages where each new file begins in the thumbnails, or type their numbers like 25,49,71. The list of resulting files appears under the thumbnails before you press the button, and they all download together in one ZIP. Page 1 starts the first file automatically, so you never name it.
How do I split a PDF into parts of a certain size?
Use the split by size tool rather than cutting by pages, and pick the ceiling the portal asks for from the five offered: 1, 2, 5, 10 or 25 MB. Each part is filled with the most pages that keep it under the ceiling, and every candidate part is actually written and its bytes counted, so the size shown beside a part is the size of the file you will download rather than an estimate of it.
What is the maximum PDF size I can split here?
There is no limit from our side, on size or on page count, because the file is never uploaded and the work runs inside your browser on your own device. The only limit is your device's memory, and cutting is among the lightest operations here because it copies pages rather than redrawing them. Thumbnails are drawn gradually, nearest to your screen first, so length alone does not stall the page.
How do I write a page range when splitting a PDF?
With numbers, commas and dashes: 1-3,5,8 means 1 through 3, then 5, then 8, and an empty field means the whole file. An open dash works too, so 50- means page 50 to the end. A range that runs past the last page stops there, and Arabic-Indic digits and the Arabic comma are accepted exactly as an Arabic keyboard types them.
Does splitting a PDF make the file smaller?
Each part is smaller than the original, but the parts add up to more than it. An embedded font or a repeated image that was stored once in one file is now stored once in every part; in a measurement here, a six-page file cut in half added about a third to the total. If what you need is a smaller document, compress it; if what you need is to get under an upload limit, splitting is the answer, because it costs you nothing in quality.
Can I split a large PDF on a phone?
Yes, the same way and with no app to install. Open the page in your phone's browser, drop the file in or pick it from the Files app or a chat, and tap the thumbnails to mark where each part begins. The result arrives as a ZIP that the Files app on iPhone and Android opens with one tap. Once the page has loaded once, the tool keeps working with no internet.
Are my files uploaded when I split them?
No. Reading, cutting and writing all happen inside your browser, the file never leaves your device at any step, and we never see what is in it. That is why this route suits contracts, bank statements and large government files, and you can disconnect from the internet after the page loads and the split still finishes.
