Rotate PDF Pages

Fix sideways scans without re-rendering the document. Rotation is metadata-only — the bytes underneath do not change.

In your browser—your files never leave your device.

Files never leave your browser. Rotation only changes page-orientation metadata — content streams are untouched (no quality loss).

Half the scanned PDFs in the world have at least one page rotated the wrong way, usually because someone fed the document into the scanner sideways or held the phone upside down. This page rotates either every page by the same angle, or only the specific pages you list (try "1-3, 5"), by 90, 180, or 270 degrees clockwise. Rotation is lossless: the underlying content is untouched, and only the page rotation flag in the PDF dictionary is updated. The file size after rotation is essentially identical to the original. Runs in your browser using pdf-lib, no upload, takes about a second even on 200-page documents because no rendering happens. The how-to and FAQ below cover the per-page mode, what happens with mixed-orientation source PDFs, and why no online tool offers true page mirroring.

How to rotate pdf pages

Drop in your PDF

Single file. The page reads the page count immediately so you can confirm what you uploaded and know the valid range for specific-pages mode.
Pick an angle

90, 180, or 270 degrees clockwise. 90 is what you want for a sideways scan; 180 for upside-down; 270 for sideways the other way. PDF rotation is always in 90-degree increments — arbitrary angles need a different tool.
Pick a scope

"All pages" rotates every page by the same angle. "Specific pages" takes a range like "1-3, 5, 8" — only those pages get rotated, the rest stay put.
Click Rotate and download

Output is a new PDF with the rotation flag updated. File size is within a few percent of the original because no content is re-rendered.

Why use this tool

You scanned a stack of mixed-orientation documents and pages 3 and 7 came out sideways. You imported a phone-camera shot into a PDF and the photo orientation got lost so the page is rotated 90 degrees. You exported a slide deck to PDF in landscape but you wanted portrait. You inherited a scanned book where every odd page is rotated 180 from every even page (whoever scanned it was flipping the book wrong). You received a contract where someone scanned page 4 upside down. You photographed a paper document and want to fix the orientation before sharing it. Rotation is one of those operations that desktop PDF readers can preview but most cannot save — Mac Preview lets you rotate but not always save reliably across versions, and Adobe Reader (the free one) explicitly cannot save rotated copies. So people end up paying for Acrobat just for this. Here it is free, lossless, and runs locally.

Features

Per-page or all-pages rotation

Pick "All pages" for the common case (one bad scan, every page sideways). Pick "Specific pages" and type something like "1, 3-5, 9" when only certain pages are wrong. This is the feature most online rotate tools skip — they only do all-or-nothing, forcing you to split, rotate, merge separately. Having the per-page mode here means a one-step fix for the typical scanner mishap where two or three pages out of fifty came out the wrong way.

Truly lossless

PDFs store page rotation as metadata in the page dictionary, not as actual transformed pixels. This tool flips that metadata flag and re-saves the PDF — no rasterization, no re-encoding, no quality loss. Text stays selectable, vector graphics stay sharp, file size barely changes. Some online rotators secretly re-rasterize to "normalize" the output, which makes text fuzzy and balloons file sizes. This one does not.

Reads page count up front

As soon as you drop in the file, the tool parses the structure and shows the page count. Two benefits: you confirm the right file uploaded, and you know the valid range for the "specific pages" field. So you do not type "1-100" on a 30-page document and get a cryptic error. If the input is encrypted, the parser flags it at this step before you waste time picking settings.

In-browser, no upload

Rotation runs through pdf-lib in your browser. The PDF never touches a server. This matters less for rotation than for compression (since rotation just shuffles metadata bits and the content never leaves the document) but the local execution model means the page works offline once loaded, there is no rate-limit, no queue, no daily cap. For documents you cannot upload (NDA-bound, regulated content), local rotation is the only safe option.

Privacy & security

This tool runs entirely in your browser. Your files are never uploaded to a server — every step of the process (reading, transforming, downloading) happens on your device using JavaScript and the Web APIs. You can verify this in your browser's network tab: clicking the tool's main action triggers zero requests to our servers. The page itself is served over HTTPS, but once it loads, your data stays put. No accounts, no tracking of file contents, no scanning your inputs.

Frequently asked questions

Can I rotate only some pages and not others?

Yes — pick "Specific pages" and type the page numbers or ranges, like "1-3, 5, 8-10". Only those pages get rotated; everything else stays exactly as it was. This is the fix you want when one or two pages in a long scan came out sideways and you do not want to mess with the rest. The range syntax matches the split tool, so once you learn one you know both.

Will the file size change?

Only by a tiny amount. pdf-lib rewrites the PDF structure during save and some object streams get re-compressed slightly differently, but the content streams (the actual page graphics) are untouched. Expect file size deltas under 5%. If you see a big size change, something else is going on — probably the source had inefficient encoding that pdf-lib cleaned up as a side effect, which is harmless but worth noting.

Does this re-render or rasterize my PDF?

No. Rotation in PDF is a metadata operation — the page dictionary has a /Rotate entry, and this tool just updates that entry. The page contents (text, vectors, embedded images) are copied through verbatim. So selectable text stays selectable, fonts stay embedded, and there is no quality loss. Compare with the compress tool, which does rasterize, and you will see the difference: rotated output looks identical to the source at any zoom level.

What about flipping (mirroring)?

Not supported here. Mirroring a page (horizontal flip, like a mirror image) is not a standard PDF rotation — it requires re-rendering each page with a transformation matrix, which is a different operation and lossy. For most use cases (sideways scans, upside-down photos), pure 90/180/270 rotation is what you need. If you genuinely need mirroring (which is rare), you would have to rasterize the PDF to images with the PDF-to-JPG tool, flip them in an image editor, and rebuild the PDF with the JPG-to-PDF tool.

Are encrypted PDFs supported?

You need to unlock them first with the PDF unlock tool. The rotate tool refuses to operate on encrypted files because handling the password input here would conflate two security concerns. Unlock, rotate, optionally re-protect — three steps but each one is one click. All three tools run in your browser so the password stays on your device.

Does it work on a phone?

Yes, the rotate flow is light on resources because nothing is rasterized — pdf-lib just rewrites the page dictionary. Even a 500-page PDF rotates in under five seconds on a modern phone. The main concern with phones is having enough memory to hold the document, and since rotation does not rasterize anything, that ceiling is high. iOS Safari and Android Chrome both handle it without issues.

Why does the page count matter?

Because the "specific pages" field validates against the page count. If you type "1-100" on a 30-page PDF the tool tells you exactly which range was out of bounds instead of failing silently or producing nonsense output. The page count is shown next to the filename so you know the upper bound when you type ranges. It also helps confirm you uploaded the right file before you start picking angles.

Will my bookmarks survive the rotation?

Yes. Bookmarks reference pages by their position in the document, not by their visual orientation. Rotating a page does not change its position, so all bookmarks still point to the right page in the rotated output. Text-selectable PDFs also keep their selectability and you can still search the document afterward — rotation does not break the underlying text layer.

Can I rotate by a custom angle like 45 degrees?

No, and no online PDF rotator can. The PDF specification only supports rotation in 90-degree increments (0, 90, 180, 270) as page-dictionary metadata. Arbitrary angles would require re-rendering each page with a transformation matrix and embedding the result as new content, which is a different operation and lossy. For artistic effects like a 45-degree tilt, rasterize the PDF to images first with the PDF-to-JPG tool, rotate the images, and rebuild the PDF.

Will it work if some pages are already rotated different ways?

Yes. Mixed-orientation PDFs (some pages portrait, some landscape, some rotated 90 from import) rotate cleanly. The tool adds the rotation angle you pick to whatever the page already has, modulo 360. So if a page was already at 90 degrees and you apply another 90, it ends up at 180. If you want every page set to a specific absolute angle regardless of what they were before, you would need to rotate twice (first set everything to a known state, then apply the target angle) or use a different tool.

What is the difference between rotating in a PDF reader vs. rotating here?

Most PDF readers (Preview, Adobe Reader free version, Foxit) let you rotate the view of a page for display purposes, but saving that rotation back to the file is not always supported. Acrobat Pro can save rotations. Mac Preview can save them but the result is sometimes inconsistent across versions. This tool produces a real saved-rotation PDF that any reader will open at the new orientation — no "view-only rotation" ambiguity.

Will rotated text still be searchable in PDF readers?

Yes. The text layer is independent of the page rotation flag — text content stays selectable and searchable after rotation. PDF readers handle the rotation when displaying the text, so a search for a word inside a rotated page still finds and highlights it correctly. This is one of the benefits of metadata-only rotation: nothing about the underlying text changes, so all the text-based features (search, copy, screen reader) continue to work.

Why does the page sometimes show 0 page count for valid PDFs?

Usually because the PDF is encrypted (the parser cannot read past the encryption layer to count pages) or malformed (the file claims to be a PDF but has corrupted structure). For encrypted files, unlock with the unlock tool first. For malformed files, try opening them in Acrobat or Preview first to see if there is a salvageable version, or use a PDF repair tool. A 0-page-count PDF is a signal that something deeper is wrong.

Can I rotate a single page out of a large PDF?

Yes — use "Specific pages" mode and type just that page number. Only that page rotates; the rest of the document is left untouched. This is the most common use case (one bad scan in a long document) and it is what the per-page mode was built for.

About this tool