Free PDF Watermark Online

How It Works

1. Upload a PDF file using the drop zone above.
2. Customize the watermark text, size, color, and opacity.
3. Choose the position (center diagonal, top, or bottom) and which pages to apply it to.
4. Click "Apply Watermark" to process your PDF.
5. Download the watermarked PDF instantly.

Why Add Watermarks?

Watermarks protect your intellectual property by marking documents as confidential, copyrighted, or under review. They discourage unauthorized sharing and printing while maintaining document readability. Common use cases include marking drafts, protecting trade secrets, copyright notices, and branding documents.

Watermark Options

• Text · Enter any text for your watermark (e.g., "CONFIDENTIAL", "DRAFT", company name).
• Font Size · Choose from 20pt to 100pt to match your document.
• Color · Pick any color; light gray is recommended for readability.
• Opacity · Adjust transparency (10-100%) to balance visibility and readability.
• Position · Center (45° diagonal), Top, or Bottom of pages.
• Pages · Apply to all pages, first page only, or a custom range.

Frequently Asked Questions

Is my PDF secure?

Yes. All processing happens in your browser. Your PDF never leaves your device and is not uploaded to any server.

Can I remove a watermark?

Watermarks added by this tool can be removed by re-processing the original PDF without a watermark. For encrypted or protected PDFs, removal may require additional tools.

What file size limit is there?

PDFs up to 100 MB can be processed. Processing time depends on file size and your device capabilities.

Can I use images or logos instead of text?

This tool adds text watermarks. For image/logo watermarks, use a desktop PDF editor like Adobe Acrobat or free alternatives.

What "PDF watermark" actually means

In everyday usage "watermark" covers at least three genuinely different things, and they are often conflated. Pulling them apart matters because the choice of technique determines what the watermark is for and how easily it can be removed. The first is a visible page-content watermark: text or an image drawn into the page itself, on top of (or behind) the document content. A CONFIDENTIAL or DRAFT stamp diagonally across every page, a company logo faintly visible in the background, a "DO NOT COPY" warning in the footer of a contract. The watermark becomes part of the page imagery; it is rendered every time the page is rendered, prints whenever the page prints, and nothing in the PDF file's structure distinguishes it from the original page content. This tool produces watermarks of this kind.

The second is a watermark annotation, called a "stamp annotation" in the PDF specification. Annotations are separate objects inside the PDF that overlay the page at display time. They look the same as content-stream watermarks when rendered, but the underlying bytes are an annotation object: a user with a PDF editor can select and delete a stamp annotation in seconds, whereas a content-stream watermark requires re-editing the page content stream itself, which most consumer tools cannot do. For visible "confidential" or "draft" marking, content-stream watermarks are the standard choice because they survive normal-user editing.

The third is an invisible watermark, also called steganographic or forensic watermarking. Information is embedded in the document in a way that is undetectable to a human reader, typically by perturbing the least-significant bits of embedded images, or by minute spacing changes between letters. Forensic watermarks typically encode a per-recipient identifier so that a leaked copy can be traced back to the original recipient. This is a specialised field used by film studios for screener copies and by some enterprise rights-management products. It is not what consumer-facing online "watermark a PDF" tools mean, and it is not what this tool produces. This tool does the first thing: it draws visible text into the page content stream of every page (or a selected page range), using a font you can size, a colour you can pick, an opacity you can dial, and a position (centre diagonal, top, or bottom).

A short history of watermarks

The word "watermark" predates digital documents by seven centuries. The earliest paper watermarks were made in Fabriano, Italy, around 1282. Fabriano's papermakers had figured out that pressing a thin wire design into wet paper pulp left a translucent pattern in the finished sheet, visible when the paper was held up to light. The original purpose was a trademark: a small wire shape (a horseshoe, a crown, a cross) identified which mill had produced the paper, useful when paper was a luxury good and quality varied widely. The technique spread to other European mills through the 1300s and 1400s.

By the late 17th century, watermarks had been adopted for a second purpose: anti-counterfeiting on banknotes. Embedding a watermark in the paper substrate made banknotes harder to fake because ordinary printing could not reproduce the translucent effect. The 1801 introduction of wavy-line watermarks (William Brewer's idea) made the technique even harder to imitate, and the 1826 invention of the dandy roll by John Marshall made watermark production efficient enough for mass currency printing. By 1848, cylinder-mould watermark paper had become the standard substrate for banknotes, passports, motor-vehicle titles, and other security documents. Modern banknotes still use this technology; the watermark you can see when you hold a 20-euro or 20-dollar note up to the light is direct continuity with the 1282 Fabriano technique. Digital watermarking borrowed the metaphor and the term, then diverged into the visible-overlay style (what this tool does) and the invisible-encoding style (a separate specialty product line).

How this tool works

A PDF page is, at the lowest level, a sequence of drawing operations stored in a content stream. The page tells the viewer: set the current transformation matrix, set the font and size, set the colour, then show this glyph at this position; move, show the next glyph; and so on. Drawing a watermark means adding more drawing operations to this stream: set the watermark font, set the watermark colour, set the alpha (transparency), set the rotation, then show the watermark text. In PDF, transparency is part of the graphics state, controlled by an ExtGState (extended graphics state) dictionary referenced by name with the gs operator. Two alpha values exist: CA controls the alpha for stroke operations and ca controls the alpha for fill operations. For filled text, which is how text is normally rendered, the ca value is the one that matters; a value of 0.3 means 30 percent opacity, the same number you set with the opacity slider in this tool.

Rotation is a transformation of the current transformation matrix (CTM), a 3-by-3 matrix that maps the page's user coordinate system to the underlying device coordinate system. The diagonal watermark this tool produces uses a -45 degree rotation, which runs from lower-left to upper-right on a standard A4 or Letter page. The library this tool uses, pdf-lib, is an MIT-licensed JavaScript PDF library by Andrew Dillon, originally released in 2018 and widely adopted for browser-side PDF generation. pdf-lib's drawText method takes a text string and an options object that maps directly onto the PDF imaging operators: size sets the font size, color sets the fill colour, opacity sets the graphics-state ca value (the library emits an ExtGState under the hood), rotate sets the CTM rotation, and x and y set the text position in user-space points (1 point = 1/72 inch).

Step by step, when you click Apply Watermark: the browser File API reads the PDF bytes into memory locally; pdf-lib parses the byte array into an in-memory document tree (pages, page resources, fonts, images, content streams); the tool creates a new empty document and copies each source page into it page by page; for each page in the apply-to range, the tool calls drawText with your chosen size, colour, opacity, rotation, and position, and pdf-lib appends the corresponding operators to the page's content stream; the tool calls save to serialise the modified document back to a byte array, then wraps the bytes in a Blob and hands them to the browser's download API. No network request fires during any step. You can verify this by opening the browser developer tools to the Network tab before clicking Apply Watermark.

Position, font, colour, opacity

Centre diagonal, the default in this tool, is the position chosen by most consumer-facing watermark tools for CONFIDENTIAL or DRAFT stamps because it crosses the maximum amount of page content. Cropping out the watermark by trimming the page is ineffective when the watermark crosses the centre at 45 degrees, because trimming far enough to remove it also removes most of the page content. This is the historical reason banknote watermarks sit in the centre of the bill rather than the corner. Top or bottom watermarks are a different design language: they function more like a header or footer, less obtrusive, less disruptive to reading, and visually marking the document without fighting it for attention. Common choices for top or bottom watermarks include a company name, a document version number, a date, or a "Page X of Y" pattern.

Font size has a working range from about 20 points (small, suited to a corner or header use) to 100 points (large, suited to a centre-diagonal banner across a whole page). The default of 50 points fills most of the visible width of a centre-diagonal watermark on an A4 or Letter page without spilling off the edges. Light grey is the default colour and the standard choice in print-publishing convention; a neutral grey works against both white-page backgrounds and the typical mix of dark text and embedded images. Red is the conventional choice for urgent markings (CONFIDENTIAL, DO NOT COPY, INVALID); blue is sometimes chosen for branded marks; black with low opacity functions as grey. Opacity is the single most important setting:

• 10 to 15 percent, subtle branding. The watermark is visible to someone looking for it and invisible to someone skimming. Used for company-logo watermarks on client-facing documents where the watermark is informational rather than restrictive.
• 15 to 30 percent, the deterrent stamp range. CONFIDENTIAL, DRAFT, DO NOT COPY are typically rendered in this range. The default in this tool is 30 percent because the most common use is a deterrent stamp.
• 30 to 50 percent, assertive marking. Reduces document readability noticeably; appropriate for sample or specimen copies where partial illegibility is part of the point.
• Above 50 percent, dominant watermark. Body text underneath becomes difficult to read. Use sparingly, and only when restricting readability is intentional.

The Web Content Accessibility Guidelines define a 4.5:1 contrast ratio between body text and background as the minimum threshold for normal text. A watermark applied with too dark a colour or too high an opacity can pull this ratio below the threshold, breaking accessibility for low-vision readers. The 15 to 30 percent range in light grey on a white background keeps body-text contrast within the accessible range.

Real-world workflows that drive watermarking

• CONFIDENTIAL and DO NOT DISTRIBUTE marking. A document is shared with external counsel, with a board, with a regulator, with a counterparty in a transaction, and the sender wants every page to carry a visual indication that the document is not for further distribution. CONFIDENTIAL is the most common single word; INTERNAL USE ONLY, NOT FOR DISTRIBUTION, ATTORNEY-CLIENT PRIVILEGED, and CLIENT NAME ONLY appear in legal, financial, and government contexts. ISO 27001 Annex A.5.13 requires that classified material be visibly marked.
• DRAFT marking during review cycles. A document is circulated for review before it is finalised. Marking every page DRAFT prevents the in-progress copy from being mistaken for a finished deliverable and signals to reviewers that comments are welcome. The watermark is removed once the document is finalised by reapplying the original unwatermarked source. DRAFT, PRELIMINARY, FOR REVIEW, and WORKING COPY are the standard choices.
• Copyright and branding watermarks. A PDF is published or shared and the author wants every page to identify the source. Company name, document title, copyright notice, URL, or author name. Light, low-opacity, often diagonal or in the margin, less about deterring use and more about preserving attribution. The closest digital analogue to the original Fabriano papermaker trademarks.
• Document version stamping. A PDF is part of a versioned series and the author wants the version visible on every page so a printed copy is unambiguously identifiable. Version 1.0, Revision 3, Final, Approved, Q2 2026. Typically a corner or header watermark rather than centre-diagonal, because the goal is identification rather than deterrent.
• Sample, specimen, and demo copies. A document is shared as a preview before a sale, and the author wants the preview to be useful enough to demonstrate value but distinct from the paid product. SAMPLE, SPECIMEN, PREVIEW, NOT FOR PUBLICATION. Higher-opacity diagonal watermarks are intentional here; partial illegibility is part of the design.
• Compliance and regulatory marking. Some regulators require specific markings on materials submitted or distributed under their rules. CLINICAL TRIAL, INVESTIGATIONAL USE ONLY, FOR INTERNAL USE - DO NOT REDISTRIBUTE, PROVISIONAL DOCUMENT. Wording and visibility requirements are prescribed by the regulator; the tool's role is to make application of the prescribed mark efficient.

Common pitfalls and what they mean

• The watermark is invisible after applying. Most common cause: opacity below 10 percent or a colour very close to the page background. Try 25 to 30 percent opacity in light grey on a white page as a known-good baseline. Second cause: the watermark text is empty or contains only spaces. Third cause: the apply-to range does not include the page you are viewing; check that "All Pages" is selected or that the custom range covers the relevant pages.
• The watermark covers the page content too much. Opacity is too high. Drop to the 15 to 30 percent range. For dark-background slide PDFs (PowerPoint exports with dark themes), the standard light-grey watermark may disappear because grey on dark is similar to grey on white in reverse; consider light text on dark documents, or accept the visibility difference between page types.
• The watermark is hard to see over images. Opacity is local to the watermark, not the page; an image with a busy multicolour background may eat the watermark visually even when the watermark is technically painted on top of it. Increase opacity to 35 to 40 percent for image-heavy pages, or move from centre-diagonal to the page corners where the background is more often uniform.
• A signed PDF becomes invalid after watermarking. By design. Adding a watermark rewrites the page content stream, which changes the document bytes, which invalidates any cryptographic signature attached to the original byte sequence. Watermark before signing, not after. If a signed document needs watermarking, the signature must be reapplied afterwards using whatever signing tool the workflow uses.
• The watermark is removable. True, with varying difficulty. A content-stream text watermark like the one this tool produces is harder to remove than a stamp annotation (which can be deleted in one click in a PDF editor), but it is not unremovable: a determined user with Adobe Acrobat Pro or a content-stream editor can rewrite the page content and delete the watermark operators. For deterrent purposes this is acceptable because the watermark exists to influence behaviour, not to make removal impossible. For piracy deterrence on copyrighted material, visible watermarks are deterrents not enforcement mechanisms; forensic invisible watermarks are the technical answer to piracy.
• The watermark looks fine on screen but prints poorly. Many office printers handle PDF transparency by rasterising the transparent layer. A 15-percent-opacity grey watermark may print as either a clean grey overlay (modern PostScript Level 3 printers) or as a jagged anti-alias-failed mess (older inkjets, some cheap laser printers). If print quality matters, test-print a representative page before applying the watermark to a large batch.
• Watermark is not a substitute for redaction. Putting a black rectangle or dense overlay on top of sensitive text is not redaction. The underlying text is still in the PDF content stream and is retrievable with any PDF text extractor. True redaction removes the bytes; a watermark or visual cover does not. If the goal is hiding text rather than marking the document, use a dedicated redaction tool.
• Non-Latin characters render as boxes. The standard Helvetica font used by pdf-lib covers Latin alphabets, common punctuation, and digits. Non-Latin scripts (Chinese, Japanese, Korean, Arabic, Devanagari, Hebrew, Cyrillic outside Latin transliteration) require font embedding that the standard pdf-lib build does not provide. Watermark text in these scripts may appear as boxes or be silently skipped; use Latin transliteration (for example "JIMI" rather than the Japanese script) as a workaround, or a desktop PDF editor for true non-Latin watermarks.

Browser-side versus cloud watermarking

The cloud PDF-watermark services in the top of search results (Smallpdf, ILovePDF, Sejda, PDF24 web, Adobe Acrobat Online) all upload your file to their servers, watermark it server-side, and serve the watermarked copy as a download. The major operators publish privacy policies committing to deletion within hours (Smallpdf one hour, ILovePDF two hours, Sejda two hours) and TLS in transit, and they have strong commercial reasons to comply. But "deletion within hours" is not "never seen," and uploading confidential documents (which is the most common reason to watermark a PDF in the first place) to a third-party server has an obvious tension with the purpose of marking the document confidential.

This tool does not upload the file. The pdf-lib library runs in your browser tab, the File API reads the PDF bytes locally, all the drawing operations happen in JavaScript inside the tab, and the Blob with the watermarked bytes is handed to the browser's download API. You can prove it: open the browser developer tools to the Network tab before clicking Apply Watermark, run the operation, and observe that no request fires with your file content. The trade-off is feature scope. Cloud services often offer image watermarks (logos), drag-and-drop visual positioning, and batch processing across multiple PDFs at once. This tool does text watermarks only, with numeric positioning, single-file processing, and up to 100 MB per file. For image or logo watermarks, a desktop PDF editor (Adobe Acrobat, LibreOffice Draw, Foxit) is the realistic option.

More frequently asked questions

What is the difference between a watermark and a stamp annotation?

A stamp annotation is a separate annotation object inside the PDF; it overlays the page at display time and can be deleted in one click by a user with a PDF editor. A watermark (what this tool produces) is drawn into the page content stream itself; removing it requires rewriting the stream, which most consumer tools cannot do. Watermarks are stronger for deterrent purposes; stamp annotations are easier to apply and remove for collaborative review.

Is this a forensic or invisible watermark?

No. Forensic watermarks (also called fingerprinting) encode an invisible per-recipient identifier in the document so a leaked copy can be traced. They are produced by specialised digital-rights-management products (Imatag, Digimarc, vendor-specific DRM systems). This tool produces visible deterrent watermarks suitable for marking documents CONFIDENTIAL, DRAFT, or similar; that is a different and more common purpose.

Can I watermark a password-protected PDF directly?

Not directly. The PDF must be unlocked first so pdf-lib can read its content streams. Use the PDF Unlock tool with the open password to produce an unlocked copy, then watermark the unlocked copy. If you want the result to remain password-protected, run the watermarked output through the PDF Protect tool to apply a new password with AES-256.

Is watermark removal legal?

Removing a watermark from a document you own or have explicit permission to modify is legal. Removing watermarks from copyrighted material you do not own, or from documents marked confidential by their owner, is generally illegal under copyright statute in most jurisdictions and may implicate other laws (trade secrets, contract breach, anti-circumvention). Consider whether you have the rights to the document before attempting removal.

Why -45 degree rotation for the diagonal watermark?

It is a convention. A -45 degree rotation runs the watermark from lower-left to upper-right on a standard A4 or Letter page, crossing both the body text and embedded images. A +45 degree rotation works equally well; both choices are equivalent for visibility and resistance to cropping. -45 is slightly more common in commercial PDF tools and is what this tool emits.

Can I batch-watermark many PDFs at once?

Not in this tool. Each PDF must be watermarked through the upload-configure-apply flow. For batch watermarking at scale, a desktop tool with a command-line interface (qpdf, pdfcpu, pdftk) or a PDF SDK is the practical answer. The reason this tool does not batch is browser memory: opening, watermarking, and saving a 100 MB PDF allocates several hundred megabytes, and running ten in parallel can run a tab out of memory.

Is there a desktop or command-line equivalent?

qpdf has an --overlay mode that overlays a one-page watermark PDF onto every page of a target document; pdfcpu has a stamp command with similar function. Both run on every platform and are scriptable for batch use. Adobe Acrobat Pro, Foxit PDF Editor, and LibreOffice Draw all expose watermark functionality through their GUIs and support image watermarks that this tool does not. For one-off watermarking the browser is the fastest option; for repeating workflows, the desktop tools win on automation.

Related Tools