JPG to PDF: How to Turn Photos and Scans into One PDF

You have a folder of photos that needs to be one PDF. A passport photo for a visa application. A scan of an ID card, a PAN card, a driving licence. Four pages of a tenancy agreement photographed on the kitchen table because the scanner is at the office. A receipt for an expense claim. The form asks for "a PDF" and you have JPGs.

Two things matter here and neither is obvious. The first is that these are exactly the files you would rather not hand to a stranger's server — a passport photo and a bank statement are the two documents identity theft is built from. The second is that converting an image to a PDF is usually not a conversion at all: it is a container operation, and for a JPEG it can be done without touching a single pixel. This guide shows the measurement that proves it, which of the tools that rank for this actually keep your file on your own computer, and the one failure everyone hits afterwards — the sideways photo.

What actually happens when a JPEG becomes a PDF

This is the part nobody explains, and it is the reason "does converting to PDF lose quality?" has a different answer for every format.

A PDF is a container. It can hold an image in one of two ways: it can carry the original compressed bytes and label them with a filter, or it can store the decoded pixels and compress them itself. A JPEG is already compressed, so the first route is available — the PDF writer copies the file in and writes /DCTDecode next to it, which tells any reader "the bytes that follow are a JPEG, decode them as one". Nothing is decoded, nothing is re-compressed, nothing is re-quantised.

I checked that rather than assuming it. A 6,418-byte JPEG went in and the PDF came out containing a 6,418-byte JPEG stream that compared byte-for-byte identical to the input. Not "similar", not "the same size" — identical, checked with a byte comparison rather than a size comparison, because two different JPEG encodings can land on the same length by coincidence and a length check would have called that a pass.

PNG works differently, and it is worth being precise about it. PDF has no PNG filter. The writer decodes the PNG and re-deflates the exact pixel samples, so the result is lossless but not verbatim — every colour value is preserved, but the file has genuinely been rebuilt. Measured: a 600 × 800 PNG came out as a 1,440,000-byte stream inside the PDF, which is exactly 600 × 800 × 3 bytes of raw RGB. The pixels are all there, uncompressed and then recompressed losslessly. If you were hoping the original PNG bytes were copied through, they were not, and a tool that tells you otherwise is telling you something it has not checked.

WebP is the odd one out. PDF has no WebP filter either, and unlike PNG the decoded result has to be re-encoded as something a PDF can hold — so a WebP image is the one format that genuinely gets re-compressed. In a three-image test (a JPEG, a PNG and a WebP) the tool reported exactly 1 of 3 images re-encoded, and the one was the WebP. That is worth knowing before you convert a folder of WebP screenshots and wonder why the PDF is smaller than the sum of its parts.

So the honest summary of image quality: JPEG is copied untouched, PNG is preserved losslessly, WebP is re-encoded. If you are converting photos from a phone, you are in the first case and the conversion costs you nothing at all.

How to turn images into one PDF in the browser

The route that never uploads anything is the Image to PDF tool. Nothing to install, no account, no queue.

1. Add your images. Drag a whole folder in or use the file picker. JPG, PNG and WebP are accepted; each row shows the file name, its pixel dimensions and its size, so you can see what you actually selected before you commit to anything.
2. Put them in page order. The rows become the pages in the order they appear. Use the up and down arrows to move a row, and the Rotate 90° button on a row if one image is the wrong way up — the button shows the current angle so you can see which way it is pointing.
3. Choose the page size.Fit to image makes every page exactly the size of its picture, with no white border and no scaling. A4, US Letter and US Legal are there for anything that will be printed or submitted to a form that expects paper. Auto, portrait and landscape control which way round a fixed page goes, and the margin slider keeps the image off the edge.
4. Create the PDF and download it. One image gives you a one-page PDF named after the image; several give you a multi-page PDF. The result line tells you how many pages it wrote, how big the file is, and how many images had to be re-encoded.

Everything happens in your browser tab. There is no server step, no upload, no temporary file on someone else's disk.

Which of these tools actually keeps your file on your computer

Search "jpg to pdf" and you will land on one of three tools. All three are competent. Only one of them is client-side, and the other two say so on their own pages if you read carefully — which is worth doing, because the use cases Adobe names are the exact documents you least want to send anywhere.

• Adobe says it on the page, in the sentence under the drop zone: "Your file will be securely handled by Adobe servers and deleted unless you sign in to save it." The same page names the use cases — a section on converting a passport-size photo, one on PAN and Aadhaar cards, one on bank statements and transaction records — and then uploads exactly those. Its own FAQ answers "How do I merge multiple JPG files into one PDF?" by telling you to convert each file separately and then sign up for a seven-day Acrobat Pro trial to combine them.
• Smallpdf contradicts itself inside a single page. "Our JPEG to PDF converter runs entirely in your browser" sits next to "our JPEG to PDF converter operates in the cloud", "files are automatically deleted from our servers one hour after the conversion", and "convert for free with a basic account". The free tier is an account, and the file waits on their disk for an hour.
• PDFGadget is genuinely client-side, and limits itself to JPG/JPEG — every other format gets sent to a second page.

You do not have to take anyone's word for which category a tool is in, including this one. Load the page, turn off your Wi-Fi, and convert a file. A browser-side tool keeps working because there is no server for it to talk to. A server-side tool fails, or refuses, or spins. That test takes ten seconds and it is the only privacy claim that cannot be faked.

Why your photo came out sideways

This is the failure that sends people hunting for a different tool, and it is not a sign that the tool is cheap.

Your phone does not store the photo the way you see it. Hold the phone upright and it records a landscape image, then writes an EXIF tag into the file saying "for display, rotate this 90° clockwise". Your phone's gallery, your laptop's preview and every photo app honour that instruction, so the picture looks upright everywhere you have ever looked at it.

A PDF is not a photo app. If a tool copies the JPEG's bytes straight into the PDF — which is the correct and lossless thing to do — the PDF has no idea about the EXIF tag, and the picture lies on its side. Some tools then compensate by rotating the whole page instead of the picture, which turns the page and its margins together and leaves you with a landscape page holding a sideways photo and a strip of white space.

The fix is to read the tag before deciding anything: if the orientation says the image needs turning, decode it, turn it upright on a canvas, and size the page from the upright dimensions. Measured on a real photo stored 800 × 600 with Orientation 6 — the classic "shot on a phone held upright" case — the row reports the displayed size as 600 × 800 rather than the stored 800 × 600, the page comes out portrait 600 × 800, and the picture inside it is genuinely upright rather than merely relabelled. That last part is checkable: the test image is red on the left half and blue on the right, and after conversion the top of the embedded picture samples as red and the bottom as blue, which is what a correct 90° turn produces. A tool that only changes a number and not the pixels fails that check.

If a picture is still wrong after that, the Rotate 90° button on its row turns it by hand, and the page follows the picture rather than the other way round.

Fit to image, A4, or a margin

Most converters give you one page size and hide it three clicks deep. The choice matters more than it sounds.

Fit to image gives each page exactly the dimensions of its picture. In the three-image test the page boxes came out as 800 × 600, 600 × 800 and 400 × 300 points — three different page sizes in one PDF, each with no white border and no scaling. This is what you want for a folder of photos going into a PDF that will only ever be read on a screen.

A4, Letter or Legal puts the image on a real paper size. With auto orientation the page follows each image's shape: in the same test the A4 pages came out 841.89 × 595.28 for the landscape images and 595.28 × 841.89 for the portrait one, so a mixed folder does not produce three pages of the same shape with two of them padded. Use this when the PDF will be printed, or when a form specifies a page size — and use the margin slider to keep the image clear of the edge, because most printers cannot print to the very border.

Putting the pages in the right order

Page order is the thing people forget until the PDF is assembled. A visa application with the photo after the bank statement rather than before it gets sent back.

The rows are the pages, in order, so the fix is to move the row. Moving the second image above the first in a two-image test reordered the output pages correctly, and a 90° turn applied to one image alone turned its page from 400 × 300 into 300 × 400 while the untouched image kept its 800 × 600 page. Reordering and rotating are per-image, so a single folder can mix upright and sideways scans and still come out right.

Making the PDF smaller

Turning on Compress images re-encodes every image as a JPEG, which is dramatically smaller for photographs and screenshots and does re-compress the picture. Measured on a single photo: 7,354 bytes down to 4,775 — about a third smaller — from one switch.

The honest trade-off is that it is the wrong switch to leave on by default. With it off, a JPEG is embedded byte-for-byte and a PNG is stored losslessly, which for photos is usually smaller than a re-encode would be, because the camera already did a good job of compressing them. Turn compression on when the PDF has to fit an upload limit, and leave it off when the document is the only copy of something.

The files it refuses, and why

A tool is judged by what it does with input that is not a clean JPEG. What this one does, so you can check it against your own files:

• A GIF is refused by name. Feed it a GIF and the row reads Not a JPG, PNG or WebP file, the summary counts it as 1 skipped, and the Create PDF button stays disabled — rather than producing a blank or broken page. A GIF is usually an animation, and quietly turning it into a single frame is the kind of silent wrongness that is worse than a refusal.
• A HEIC photo is refused too. iPhone photos are HEIC by default and PDF has no way to hold one. Convert it with the HEIC to JPG tool first, then drop the JPG in.
• A 16-bit PNG still works. Deep-colour PNGs are a format the browser's JPEG encoder cannot handle, so the tool falls back to a lossless path rather than failing: a 300 × 200 16-bit PNG produced a valid one-page PDF at exactly 300 × 200, reported as every image embedded untouched.
• A Word document renamed to .jpg is refused, because the tool reads the file's actual bytes rather than trusting its extension. A file with the wrong extension is the most common cause of "it just produced an empty PDF".
• Your images are not resized or cropped. Nothing is scaled down to fit a page, and nothing is centre-cropped to make a tidy shape. A 4000-pixel photo stays a 4000-pixel photo; only the page box changes around it.

The short version

Converting images to a PDF is a container operation, not a re-compression. A JPEG goes in byte-for-byte and comes out identical, a PNG is preserved losslessly, and a WebP is the one format that has to be decoded. If you are making a PDF out of phone photos, the conversion costs you nothing.

Do it in a tool that runs in your browser rather than on a server — a passport photo and a bank statement are the documents you least want sitting in someone else's temporary folder, and the way to check any tool's claim is to turn off your Wi-Fi and try it. Put the pages in order before you build, watch for photos that come out sideways (that is an EXIF tag the tool ignored, not a bad photo), and leave compression off unless you are fighting an upload limit.

Turn your images into one PDF in the browser — any number of pages, any order, and nothing leaves your computer.