Photos as a PDF, not six attachments
Somebody asks you for pictures of an item, or a supplier asks for the damage, or a form has one upload field and you have nine photos. You could send nine files. You could also send one PDF, and in most of those situations that is the better answer — not because PDF is a nicer format, but because it fixes the three things that go wrong with loose image files.
What loose image files get wrong
The order is not part of the file. Nine attachments arrive as nine icons, sorted by name or by date or by whatever the recipient's mail client decided. If the photos are steps in a sequence, or angles of the same object that only make sense in order, that ordering lives in the sender's head and nowhere else. A screenshot of the conversation may be the only thing holding it together.
They get recompressed on the way. Some messaging apps reduce images hard, some resize them, and the person who asked for "a clear photo of the scratch" gets a 800-pixel version where the scratch is a grey smudge. A PDF is passed through untouched by most systems, because it is not recognised as an image to optimise.
There is nowhere to put a caption. "The second one is the underside, ignore the shadow in the third" has to go somewhere, and it ends up in a message that scrolls away. In a PDF it sits on the page next to the photo, or on a cover page.
Why most converters cost you quality
Almost every free image-to-PDF tool works the same way: it decodes the image into raw pixels, draws those pixels onto a page canvas, then encodes the canvas again as an image to store inside the PDF. Two full compression passes, one on each side.
For a screenshot of text, that second pass is where the letters start to blur. For a product photo, it is where the fine texture on fabric goes waxy. And it happens even when the source is already a JPEG, which is what makes it feel gratuitous: the file was already compressed once, competently, by the camera.
PDF has a filter called /DCTDecode, and it exists precisely so an already-compressed JPEG can be stored without decode-and-re-encode. Every PDF reader knows how to expand it, because it is the same maths a JPEG decoder does. So the file goes in whole, and the width and height are read out of its own header.
| Source | What happens in the Images to PDF tool |
|---|---|
| JPEG, upright | Copied in whole. Zero loss, zero recompression. |
| JPEG, stored rotated | Redrawn upright once at quality 92, because a PDF page has no orientation tag |
| PNG | Converted to JPEG at 92, because /DCTDecode only reads JPEG |
| WebP | Same as PNG |
This is written on the tool page per file before you build anything, which is the point. A converter claiming to be lossless for every input is either not converting PNGs or not mentioning what it did.
The rotation problem, which nobody warns you about
A phone held upright does not store an upright image. It stores the sensor's landscape frame plus a tag in the EXIF block saying "display this turned 90 degrees". Every viewer worth using reads that tag and rotates on the fly, which is why portrait photos look portrait everywhere.
PDF has no equivalent tag for an image object. So the rotation has to be resolved before the image goes in — either by drawing it turned on the page, or by baking the rotation into the pixels. The second is what this tool does, at quality 92, and only for photos whose tag is not "already upright". It is the one case where being correct costs a little quality, and it is why the tool cannot honestly call itself lossless for every possible input.
What a PDF will not give you
- Searchable text. These pages hold pictures. If you photographed a document and need the words out of it, that is OCR, and no image-to-PDF step gets you there.
- A smaller file. Removing compression also removes the saving. Twenty 4 MB photos become a 80 MB PDF, not a 5 MB one. If a size limit is the requirement, compress first with Compress to a File Size, then build the PDF.
- Per-photo captions you can edit later. The PDF is written once. To change a caption or the order, rebuild it — which takes a second, so this is not much of a limitation in practice.
- Protection. Nothing here encrypts the output. A PDF with your home in the background is as readable as the JPEG was, metadata aside — and note that a JPEG embedded as-is still carries its EXIF, including GPS, if it had it. Strip that first if it matters: Remove Photo Metadata.
Picking a page size
Three choices, and the right one depends on whether the document will be printed.
A4 portrait is the default. Every page comes out the same shape regardless of the photo's own proportions, so a stack printed at a shop does not alternate between portrait and landscape with a white band down one side. Mixed photos of a single product look like a document rather than a shoebox.
A4 matched to each photo turns the page for landscape shots. On screen it looks more deliberate; printed, it is the version people complain about.
The photo's own size throws away the idea of a page: no margins, no white space, proportions equal to the image. Good for handing photos to someone as a bundle they will open on screen, less good as a document.
Checking the result yourself
Open the PDF, zoom to 100%, and look at fine detail — small text in a screenshot, fabric texture in a product photo. Then compare against the original side by side. If the PDF looks softer, the image was re-encoded somewhere; if it looks identical, it was carried through.
For the analytically minded: the file size tells you. A PDF holding photos embedded as-is is the sum of the JPEGs plus about a kilobyte per page. A PDF that is noticeably smaller than the sum has been recompressed, and the missing bytes are detail you no longer have.
Build the PDF
The Images to PDF tool assembles the file in your browser with no upload, labels each photo as untouched or re-encoded, and keeps page sizes and order under your control.
Related: Compress to a File Size when the file must shrink, Merge PDF to combine the result with an existing document, and Remove Photo Metadata before any of it leaves your machine.
Written October 2026. Filter names follow the PDF specification (ISO 32000); the rotation tag is EXIF tag 0x0112. Figures for compression quality are the settings this tool uses, not a claim about what other tools do.