Add Image to PDF

This tool puts a picture onto pages of a PDF you already have and gives you a new file. It is for the everyday finishing job: stamping a company logo into a corner, signing off a document with a scanned signature image, dropping a "PAID" or "APPROVED" mark onto a form, or repeating a small badge across a report. You choose the image, the position, the size, the opacity and an optional rotation, and the picture is drawn only on the pages you include.

This is the opposite of JPG to PDF and PNG to PDF. Those tools build a brand-new PDF whose pages are the images — the picture becomes the page. Here the PDF comes first and stays: its pages, page count, page sizes, rotation, text, links, bookmarks and form fields are left exactly as they were, and the image is drawn on top of the chosen pages' content. So the question to ask is "what do I already have?" — if you have a document and want a logo on it, use this tool; if you have photos and want them as a PDF, use JPG to PDF or PNG to PDF.

Placement is in points (1 point = 1/72 inch) measured on the page as you see it. Nine fixed positions are offered — the four corners, the four edge midpoints and the centre — each with a margin that is the clearance from the page edge; 24 pt is about 8.5 mm. Custom lets you type an x and a y from the top-left of the visible page to the image's top-left corner. The size is either a share of the page width (a percentage) or an exact width in points, and the height always follows the image's own proportions. Opacity fades the picture (useful for a faint watermark-like mark that still leaves the text readable), and rotation turns it by any angle in whole degrees.

Two coordinate facts make the placement match what you see. The page is measured as displayed, so a page carrying a /Rotate value — common on scanned forms — is turned before anything is placed, and the image lands where the reader will look for it, not sideways. And PDF user space runs from the bottom-left with y increasing upwards, while the numbers this tool asks for run from the TOP-left with y increasing downwards, matching how every image editor and screenshot works; the conversion is done for you. A JPEG's EXIF orientation tag is honoured too, so a phone photo that viewers show upright is placed upright rather than on its side.

What "adding" means: the image is drawn into each chosen page's own content, so it is baked in — every reader and printer shows it, and it stays put. It is not an annotation you can click, select, drag or delete in a reader, and it is not a form field. Existing content is not touched: text, links, bookmarks, form fields and metadata are kept. Because the image is drawn on top, put it somewhere there is room — over a busy area it can cover text or a signature line, and at full opacity it hides whatever is beneath it.

Some cautions, stated plainly. An encrypted or password-protected PDF (including one locked only against editing) is refused with a clear message, because the library used here cannot decrypt it — open it in a PDF reader with its password, save or print an unprotected copy, and use that. Saving a PDF with pdf-lib rewrites the file, so a digital signature on the original stops validating: sign a document after you finish stamping it, or accept that any signed copy will no longer verify. Colour profiles are not managed: the image bytes are embedded as they are, but nothing is done to reconcile a CMYK or wide-gamut image with the document's colour space, so a colour-critical logo may print or display slightly differently than in a program that handles profiles. Very large PDFs and images are held in the tab's memory while they are edited. Everything runs in your browser: the two files are read, combined and saved inside the tab, and neither is ever uploaded. You can confirm in your browser's developer tools, on the Network tab, that no request carries your files.

1. The PDF's leading bytes are checked for the PDF signature, then pdf-lib opens it; a damaged, empty or encrypted file is refused with a specific message, and the page count and each page's displayed size are read. 2. The image's bytes are checked: it must really be a PNG or JPG (the magic bytes decide, not the file name), it must be within the size limit, and its pixel size and any EXIF orientation are read. 3. The placement settings are validated: a known position, a width between 1% and 100% of the page width (or 2 to 2000 pt exactly), an opacity above 0 and at most 100%, a margin between 0 and 200 pt, a rotation that is a number, and — for Custom — both an x and a y. 4. The pages to stamp are all pages or the list you typed (numbers, ranges like 2-6, and 8- for page 8 to the end), checked against the page count. Pages are processed in ascending order. 5. Each page is measured as displayed: its visible area is the /CropBox clipped to the /MediaBox, and its /Rotate is applied, so width and height swap on a turned page. The image's displayed size is the chosen width with the height scaled by the image's own proportions. 6. The image's rotation is turned into its axis-aligned bounding box — a 100 × 40 image at 90° needs a 40 × 100 slot — and that box is anchored at the chosen position with the margin, or placed at the custom x/y. The image itself is centred inside that box and turned by the rotation. A box that would run off the page is refused with a message naming the page rather than being silently clamped. 7. The embedded image is written once. On each chosen page an XObject is registered in the page's resources and a content-stream operation is appended: save the graphics state, apply the transformation (the displayed page, the rotation and the image's EXIF orientation composed into one matrix), draw the image, restore the state. When the opacity is below 100% a graphics state with the non-stroking alpha is registered per page. 8. The document is saved as <name>-with-image.pdf. Your original file is not changed.