The entire online PDF editing industry renders PDFs to Canvas — a frozen pixel buffer with no editing capability. We invented a different approach.
Patent-pending DNPR technology.
PDF.js (Mozilla) and pdfium (Google/Chromium) are the two libraries that power virtually every online PDF service you've ever used. They're excellent at what they were built for: viewing PDFs.
But here's the problem: they were never designed for editing.
Both PDF.js and pdfium follow the same rendering approach. They read the PDF binary stream, interpret it, and paint pixels onto an HTML Canvas element. The result is an image — a flat, non-interactive picture of your document.
HTML Canvas wasn't even designed for PDFs. It was created by Apple in 2004 for macOS Dashboard widgets. PDF.js and pdfium adopted it because PDF drawing commands happen to map neatly to Canvas methods — both are immediate-mode 2D drawing systems. But that mapping only works for rendering. Once pixels are painted, the document structure is gone.
Canvas is fundamentally a pixel-based drawing surface. Once the PDF is rendered to canvas, it's no different from a screenshot. There's no text you can select natively. No elements you can click. No structure you can manipulate. It's pixels all the way down.
So how do existing services offer editing? They build additional layers on top of the canvas. Text overlay layers synchronized with coordinates. Input fields positioned over the canvas image. Custom font matching. Mode switches between "view" and "edit".
This is why every online PDF editor feels heavy. Layer over layer. Mode switching. Limited graphic control. You're never actually editing the PDF — you're editing an approximation built on top of a picture of it.
The engine can extract text content and styles, but it cannot represent the PDF stream as DOM elements. That fundamental limitation cascades into every interaction.
PDFox V2 is powered by DNPR — a proprietary PDF 2.0 interpreter (patent pending) built on the latest PDF standard. Instead of painting pixels to a canvas, it reads the PDF binary stream and directly represents it as editable HTML DOM elements.
From binary to editable DOM — zero addon layers
Your PDF becomes a live, editable document — not a picture of one. Text you can click and type. Graphics you can select, move, and resize. Images you can reposition. Every element is individually addressable, supporting deletion, replacement, and repositioning.
The approach is formally designated DOM-Native PDF Rendering (DNPR) and is the subject of U.S. Provisional Patent Application #63/997,982, filed March 6, 2026. Unlike PDF.js or pdfium — which implement older PDF specifications — the DNPR engine is built on the latest PDF 2.0 standard with full compliance. That means native support for enterprise-grade encryption, digital signatures, and full accessibility support.
Powered by DNPR — proprietary PDF 2.0 interpreter (patent pending). A purpose-built engine, not a wrapper around existing libraries:
All of this runs entirely in your browser — a privacy-first, serverless architecture where no document data, page content, font binary, or user edit is transmitted to any remote server at any point during the document lifecycle.
No overlay layers. No mode switching. No addons. When you open a PDF in PDFox V2, you get a document you can edit like a Google Doc — because the rendering model was designed for it from day one.
| PDFox (DOM) | Everyone Else (Canvas) | |
|---|---|---|
| Text selection | Native, browser-level | Simulated overlay, often misaligned |
| Accessibility | Screen readers work out of the box | Requires separate text layer hacks |
| Search | Cmd+F just works | Depends on extra implementation |
| CSS styling | Full control over rendered elements | None — it's pixels on a bitmap |
| Large documents | Renders what you see, DOM-efficient | Full page rasterization per viewport |
| Privacy | Nothing leaves your machine | Architecture limits local capabilities |
| Inline editing | Native contenteditable — browser handles cursor, IME, clipboard | Requires custom text input subsystem from scratch |
| Zoom quality | Zero-cost CSS transform: scale() — infinite resolution | Must re-render at each zoom level |
| Object-level control | Select, delete, resize, reposition any element | Architecturally impossible — pixel buffer is opaque |
| Offline support | Full offline operation after initial page load | Requires server for processing |
Open any PDF and edit it directly. No tools to select, no modes to switch. Click and type. That's it.
SVG paths, vector graphics, and images are DOM elements you can select, move, resize, and modify directly.
Built on the latest PDF 2.0 standard. Enterprise-grade encryption, digital signatures, and archival compliance — not retrofitted, native.
There's no upload step to skip or toggle. Your PDF physically cannot leave your device because the entire processing pipeline runs in-browser. GDPR compliance isn't a checkbox — it's the architecture.
Text selection, find-on-page, copy-paste, accessibility — these aren't features built on top of a canvas hack. They come free because the content is real DOM.
Open PDFox in your browser. Drop a file. Edit. Done. Works on any modern desktop browser — Chrome, Firefox, Edge, Safari.
Dive deeper into the engine, architecture, and features behind PDFox V2.
Powered by DNPR — proprietary PDF 2.0 interpreter (patent pending). Full PDF 2.0 compliance with proprietary acceleration.
Learn more →Modular design with triple-output rendering (DOM + Canvas + SVG).
Learn more →Toolless editing, graphic control, PDF/A compliance, digital signatures, accessibility, and more.
Learn more →The proprietary technology powering DNPR and PDFox V2.
Learn more →All PDF processing executes exclusively within the user's browser. No document data, page content, font binary, or user edit is transmitted to any remote server at any point during the document lifecycle.
PDFox is deployable as a fully serverless application: static assets served from any CDN, requiring no application server, no database, and no network connectivity after initial delivery. Offline operation is supported in full.
Your first document is free. After that, full access is €2.99 for 24 hours — no subscriptions, no recurring charges, no account required.
Try Free