Convert HEIF to PDF

Institutions standardised on PDF because a PDF is a page: it has a size, it prints predictably, it can be stamped and archived, and it looks the same to the person receiving it as to the person sending it. An image file is none of those things, which is why the upload box says PDF and rejects everything else.

Meanwhile the picture came off a phone in a container that half of the world's software cannot open. Converting solves both problems at once, and the only decisions left are about the page rather than about the picture — which is what most of this page is about.

PDF measures everything in points, of which there are 72 to the inch. "Fit to image" makes the page exactly as large as the picture in those units, so a 4,032 by 3,024 photograph becomes a page 4,032 by 3,024 points — that is 56 inches by 42, or roughly 1.4 metres by 1.1.

On a screen this is invisible and harmless: the viewer scales the page to the window and the picture fills it edge to edge, which is exactly what you want for an attachment somebody will glance at. Send that same file to a print queue or to an institution that checks page dimensions and the answer is different, which is what the paper sizes are for.

Choosing A4 makes the page 595.28 by 841.89 points, which is 210 by 297 millimetres. US Letter is 612 by 792 points, or 8.5 by 11 inches. Both are exact, both are what a printer and a document management system expect to see, and either is the right choice when the specification names a paper size.

The picture is then centred on that page and scaled down to fit inside the margins. A landscape photograph on a portrait page leaves white above and below it; the orientation control turns the page instead if that is not what you want. The image itself is never rotated or cropped to make it fit.

Scaling runs one way only. If the photograph is larger than the room available it is reduced to fit; if it is smaller it is left at its own size and centred, with white around it.

That rule exists because the alternative looks worse in every case that matters. Blowing a 600-pixel image up to fill an A4 page turns crisp text into soft grey shapes and makes a document photograph look like a fax. A smaller picture on a clean page reads as deliberate; a stretched one reads as a mistake.

PDF can carry two image formats natively: JPEG and PNG. A JPG dropped onto this site is embedded byte for byte, costing nothing in quality and almost nothing in time. HEIF is not one of the two, so the picture is decoded to pixels and re-encoded as a PNG on the way into the document.

PNG is lossless, which means nothing further is lost and also that the photograph is stored the expensive way — a two megabyte HEIF can become a fifteen or twenty megabyte PDF. This page has no lever for that: it offers page size, orientation and margin, and none of the three changes a single pixel. If the upload has a size limit, go through JPG instead — convert the HEIF to JPG at 1,600 pixels wide, which is plenty for a document read on screen, then drop that JPG here, since a JPG is embedded byte for byte rather than re-encoded as PNG.

Each HEIF becomes a single-page PDF. Converting a folder gives you one PDF per photograph, with a progress row each and an "all as ZIP" button once two are finished, but it does not assemble them into a document.

Where the requirement is a single multi-page PDF — a tenancy application with six photographs, a claim with front and back of a receipt — convert each picture, then merge the PDFs in a tool built for that. Doing it in that order also lets you fix the page order, which is the part that goes wrong when a converter guesses at it from file names such as IMG_4471 and IMG_4472.

A PDF made this way holds an image of a document, not the document. Nothing in it can be selected, searched, copied or read by a screen reader, and a system that scans submissions for a reference number will find nothing.

That is frequently acceptable — a human opens the file and reads it. Where it is not, the HEIF to TXT page runs character recognition and hands back the words as plain text, which you can paste into the form alongside the picture. The two answers solve different halves of the same problem.

Since the PDF preserves whatever the camera captured, the quality of the submission is decided before the conversion. Lay the paper flat, shoot straight down rather than at an angle, and put the light to one side so your own shadow is not across the text.

Fill the frame with the page and nothing else. A photograph of a desk with an A4 sheet in the middle of it wastes most of its pixels on the desk, and when the PDF is scaled to fit a page the document ends up small, soft and hard to read — which is the version that gets sent back with a request to resubmit.

The photograph's EXIF does not survive the trip: no camera model, no capture timestamp, no GPS coordinates land in the document. The conversion decodes to raw pixels and re-encodes, and metadata has no path across that step.

For a submission to an institution this is a sensible default. It also has a consequence worth knowing: if the recipient expects the photograph to prove when or where something was taken, the PDF cannot support that claim, and the original HEIF is the only file that carries the evidence.

Decoding, encoding and the PDF writing all happen inside the browser tab. No copy of your document reaches a server, there is no queue and no account, and the free tier accepts inputs up to 100 MB each.

For this pair that is more than a nicety. The photographs people convert into PDFs are payslips, passports, medical letters, tenancy agreements and bank statements — files whose entire risk profile is who else has a copy. Here the answer is nobody, and the network tab is where you can check it rather than take our word.

On a fixed paper size the page starts portrait, and a landscape photograph is scaled to fit the width with white above and below it. That is often perfectly acceptable and occasionally wasteful: a wide receipt reduced to fit a portrait A4 ends up small enough that the print is hard to read.

The orientation control turns the page rather than the picture. Setting it to landscape swaps the page dimensions, so a wide photograph fills far more of the sheet. The margin setting works alongside it, holding the picture away from the edges of a fixed sheet, which is what most print and scanning processes expect to see. Orientation has no effect when the page size is fit-to-image, because there the page already has the picture's own proportions and there is nothing to letterbox.

The picture goes into the document with no ICC profile attached, so any viewer or printer reading the PDF interprets the values as ordinary sRGB. On screen that is what everybody assumes anyway and the document looks as expected.

It matters if the submission will be printed and the colours are the point — a photograph of damage for an insurance claim, a swatch, a medical image. In those cases the safest step is to look at a printed proof rather than at the screen, and to accept that a converter cannot manage colour for a press it knows nothing about. For a photograph of a document, which is the usual case here, none of this is worth a second thought.