Phone photos come out the right way up. Cameras rarely rotate the pixels — they record the orientation in a tag and leave the image on its side, which is why so many converters produce PDFs full of sideways photos. Every image here is decoded with that tag applied, and read by hand where the browser will not do it.

Turn photos, scans and screenshots into a single PDF without uploading them. Real page sizes, margins, and the option to fit two or four images to a sheet — which is what receipts and ID cards actually need.

Images to PDF

Local

Choose images, or drop them here JPEG, PNG, WebP, GIF, BMP — nothing leaves this tab

Written by pdf-lib, fetched when you add your first image. Your files are not part of that request.

The camera never turns; the phone records which way it was held. That is faster than rewriting the image, and it works perfectly until the file reaches something that does not check.

Why phone photos come out sideways

A camera held vertically does not usually rotate what it records. It stores the picture in the sensor’s own orientation and writes a note — the EXIF orientation tag — saying which way up it should be displayed. Every photo viewer reads that note. Many converters do not.

Tag value	Means	When you see it
1	Already the right way up	Screenshots, scans, anything edited and re-saved
6	Rotate 90° clockwise	Phone held upright — by far the most common
8	Rotate 90° anticlockwise	Phone held upright the other way
3	Rotate 180°	Phone upside down
2, 4, 5, 7	Mirrored, with or without rotation	Front cameras, some scanning apps

All eight are handled. Where the browser supports decoding with the orientation applied, that is used; where it does not, the tag is read out of the file’s bytes and the rotation applied to a canvas. An image whose tag was anything other than 1 is marked in the grid, so you can see the correction happened rather than hoping.

Page size, and what fits inside it

Setting	What it does
Fit the page to each image	Every page takes the exact proportions of its image. No white space, no cropping — but pages of different sizes, which prints badly
A4, Letter, Legal, A3, A5	Every page the same. Each image is scaled to fit inside the margins, keeping its proportions, and centred
Orientation	Chosen per page from the first image on it: a landscape photo gets a landscape page without you asking — for one or two images per page. A four-image sheet is always upright, since it is a fixed grid rather than a shape that follows one photo
Margin	White space around the edge, in millimetres. 10 mm suits most printers; 0 for a borderless look on screen

Use a real page size for anything that will be printed. A PDF whose pages are all different sizes confuses printer drivers and gets scaled unpredictably; the convenience of no white space is not worth it once paper is involved.

Two or four to a page

One image per page is the right answer for photographs and scanned documents. It is the wrong answer for receipts, business cards, ID photos and anything you are filing rather than reading:

Per page	Good for	Note
One	Photos, full-page scans, presentation slides	Each image gets the whole sheet
Two	Receipts, ID cards, passport pages	Stacked on a portrait page, side by side on landscape — decided by the first image's shape
Four	Expense claims, contact sheets, thumbnails	A two-by-two grid, always on a portrait sheet. Small text will not be readable

Four to a page makes documents unreadable. A page of A4 text photographed and then shrunk to a quarter of a sheet is roughly 3-point type. Use it for images you need to identify at a glance, not for anything anyone has to read.

File size, and the quality setting

Ten photos from a modern phone are 40 to 80 MB before anything is done to them. The quality setting is the only control that changes that meaningfully:

Setting	Ten phone photos	Use it when
Original	40–80 MB	Archiving, printing, anything where the image is the point
High — 92%	8–15 MB	The usual answer. Differences are hard to see at normal viewing size
Medium — 80%	4–8 MB	Email attachments, upload forms with a size cap
Small — 60%	2–4 MB	Screen reading only. Visible artefacts around text and edges

Original re-encodes nothing — JPEG and PNG files are embedded exactly as they arrived, which is both the fastest option and the only lossless one. Every other setting decodes and re-encodes as JPEG, so choosing High on an already-compressed photo loses a little quality it can never get back.

Resolution decides the printed size, not the pixels

A pixel has no physical size. What it becomes on paper depends entirely on how many are declared per inch:

A 3000 × 2000 photo at	Prints at	Suits
300 dpi	254 × 169 mm	Photographic quality. The standard for anything being printed properly
150 dpi	508 × 339 mm	Acceptable for office documents and text scans
96 dpi	794 × 529 mm	Screen resolution. Visibly soft on paper

To fill an A4 page at 300 dpi you need about 2,480 × 3,508 pixels. Anything smaller is being stretched, and the shortfall shows as softness rather than as an error message. A modern phone photo at 12 megapixels comfortably exceeds this; a screenshot almost never does.

Which format to feed it

Format	Best for	Note
JPEG	Photographs	Lossy, small, and embedded directly without re-encoding — no second generation of loss
PNG	Screenshots, diagrams, text	Lossless. Far larger for photographs, far sharper for anything with hard edges
HEIC	iPhone default	Converted before embedding. Set the camera to “Most Compatible” to avoid the step entirely
WebP	Web images	Converted before embedding. Widely supported now, but not natively by PDF
Screenshots of text	Nothing	Use PNG rather than JPEG. JPEG compression puts visible fringing around letters

A PDF made from photographs contains no text. It cannot be searched, selected, copied or read by a screen reader, however clearly the words appear. If the document needs to be searchable, it needs OCR — and a photographed page is a picture of a document, not a document.

Common mistakes

Rotating photos by hand first. If a converter shows your photos sideways, the images are almost certainly fine and the tool is ignoring the orientation tag. Rotating them in an editor and re-saving strips the tag, which fixes that tool and can break others. The tag is not the problem.

Expecting the text in a photographed document to be searchable. A photo of a page is a picture, whatever it contains. The PDF holds an image and no text at all — searching it finds nothing, and copying from it copies nothing. Making it searchable needs OCR, which is a different tool.

Uploading identity documents to a website. Passport scans, driving licences and bank statements are exactly what people convert to PDF, and exactly what should not be posted to a server whose retention policy you are taking on faith. Anything that runs in the browser removes the question rather than answering it.

Using “fit to image” for something that will be printed. Pages of varying size make a printer scale each one differently, and a document that looked fine on screen comes out with inconsistent margins. Pick A4 or Letter whenever paper is the destination.

Frequently asked questions

Why do my photos come out sideways in other tools?

Because phone cameras store the picture in the sensor’s orientation and record which way up it should appear in an EXIF tag rather than rotating the pixels. A converter that ignores that tag produces sideways pages. All eight orientation values are handled here, and any image that needed correcting is marked in the grid.

Are my images uploaded?

No. They are read by your browser and the PDF is written there too — the network panel in developer tools will show no request carrying an image. The only network request is for the library, and it happens once.

Which formats work?

JPEG and PNG are embedded directly, without re-encoding, on the Original quality setting. WebP, GIF and BMP are converted to JPEG first because PDF cannot hold them. HEIC — the default on recent iPhones — is not supported by browsers and will not open; export as JPEG from the Photos app first, or turn on Most Compatible in the camera settings.

Can I put several images on one page?

Yes — two or four. Two suits receipts and ID cards, and the page follows the shape of the first image on it: stacked on a portrait page, side by side if that image is landscape. Four suits expense claims and contact sheets and is always a portrait sheet, since it is a fixed grid rather than one that adapts to a photo. Bear in mind that a full page of text reduced to a quarter of a sheet is roughly 3-point type and will not be readable.

Will the text in my photos be searchable?

No. A photo of a page is a picture, so the PDF contains an image and no text — searching finds nothing and copying copies nothing. Making it searchable requires OCR, which is a separate technology and a separate tool.

How do I make the file smaller?

Lower the quality setting. Ten phone photos are 40 to 80 MB at Original and 8 to 15 MB at High, which is usually indistinguishable at normal viewing size. Medium suits email attachments; Small is for screen reading only and shows artefacts around text.

Can I change the order?

Drag any thumbnail to move it. The order in the grid is the order in the PDF. You can also rotate an individual image or remove it with the buttons that appear on each one.

Is there a limit on how many images?

No. Server-based tools cap free use because each conversion costs them processing time; nothing runs on a server here. The practical ceiling is your own machine’s memory — a few hundred photos is fine, though it will take a moment.

Related tools

See the full list of Document Tools, or try:

Merge PDF — combine the result with other documents
Split PDF — take it apart again
PDF to Markdown — pull text out of a PDF
Markdown to PDF — make one from written text
PDF to Images — the other direction, exporting pages as pictures

Images to PDF: Correct Rotation, Multi-Up, No Upload