Read documents
without the upload.
A clinical, privacy-first document renderer. Drop a PDF, DOCX, or EPUB file, and your browser parses it instantly. Nothing touches a server.
The Architecture of Privacy
Most online document viewers function as data honeypots. They require a server round-trip, exposing sensitive files to third-party infrastructure.
Read the Security ProtocolZero Server Uploads
Files never leave your device. The rendering engine is compiled to WebAssembly and runs entirely within your browser's sandboxed environment.
Zero Latency
Without network overhead, large PDFs and complex DOCX files parse in milliseconds. The bottleneck is your CPU, not your connection.
Ephemeral State
When you close the tab, the document ceases to exist. We utilize aggressive garbage collection and eschew local storage for file contents.
Strict Content Security
Scripts embedded within PDFs are neutralized. Our parser extracts layout and text, discarding potential execution vectors.
Supported Formats
| Format | Engine | Fidelity | Action |
|---|---|---|---|
|
PDF
Portable Document Format
.pdf
|
pdf.js (Mozilla) |
|
Specs |
|
DOCX
Microsoft Word Open XML
.docx
|
mammoth.js (Custom) |
|
Specs |
|
EPUB
Electronic Publication
.epub
|
epub.js |
|
Specs |
|
JSON
JavaScript Object Notation
.json
|
V8 Native JSON.parse |
|
Specs |
The True Cost of "Free" Document Viewers
When you upload a sensitive contract or financial record to a cloud-based viewer, you surrender control. Their business model relies on aggregating metadata and, in some cases, scanning text for advertising profiles.
-
Standard Cloud Viewer Uploads file -> Extracts text -> Stores on S3 -> Serves HTML -> Logs IP/Metadata
-
uViewDoc Protocol Select file -> Browser WebWorker parses bytes -> Renders directly to Canvas/DOM. No network request made.
Diagnostic Tools
Beyond rendering, uViewDoc provides a suite of clinical tools for document analysis and structural inspection. All run entirely in-browser.
Deep Word Counter
Extracts and counts text accurately from DOCX and PDF, ignoring footnotes or formatting artifacts.
JSON Formatter
Strict validation and formatting. Handles large payloads efficiently without crashing.
Base64 Decoder
Decode base64 strings back to text or binary files instantly on your machine.
Markdown Preview
Live compile Markdown to HTML securely within the browser sandbox.
Engine Benchmarks
PDF Initialization
Average time to parse PDF header structure and build page tree references on M1 architecture.
Network Payload
Bytes sent to our servers when you view a document. The viewer is a static HTML/JS bundle.
Scroll Performance
Virtual DOM rendering ensures steady framerates even on 500+ page technical manuals.
GARBAGE_COLLECTION_PROTOCOL
> File pointer received via ArrayBuffer.
> PDFDocumentProxy initialized.
> Rendering Page 1 to CanvasContext2D.
> ...
> User closes tab.
> Buffer dereferenced. Memory reclaimed instantly.
Handling Massive Files
Browsers impose strict memory limits (often capping at 2GB). Loading a 500MB architectural PDF directly into memory will crash most tabs.
We utilize lazy-loading streams and progressive canvas recycling to ensure stability. Only visible pages are rendered, meaning a 1,000-page document consumes roughly the same RAM as a 10-page document.
Read the Memory Limits GuideDefensive DOM Injection
Formats like DOCX and EPUB are essentially zipped XML/HTML. Rendering them means converting that XML to DOM nodes on your screen.
If a malicious actor embeds a <script> tag inside an EPUB chapter, rendering it blindly would trigger a Cross-Site Scripting (XSS) attack locally. We run all parsed flowable text through a strict DOMPurify pipeline before it ever touches the Document Object Model.
- <script src="..."> STRIPPED
- onclick="eval()" STRIPPED
- <iframe> STRIPPED
- <strong>, <em> ALLOWED
Zero-Request Font Mapping
If a document requires a font you don't have installed, we refuse to fetch it from Google Fonts or Adobe. Network requests leak metadata. Instead, we use a geometric heuristic to horizontally scale a local system font to perfectly match the original font's bounding box, preserving layout fidelity without compromising privacy.
Explore Font Substitution LogicFrequently Asked Queries
Embeddable Local API
Need to render documents securely within your own internal tools? We provide a lightweight wrapper around our core rendering loop.
Initialize the viewer by passing a File object directly from an input element. No server backend required.
Read the Architecture Docsimport { DocumentEngine } from '@uviewdoc/core';
const fileInput = document.getElementById('doc');
const canvas = document.getElementById('renderTarget');
fileInput.addEventListener('change', async (e) => {
const engine = new DocumentEngine({
sandbox: true,
disableNetwork: true
});
// Renders locally, zero HTTP requests
await engine.renderPage(e.target.files[0], canvas, 1);
});
Hardware Compatibility Matrix
Built for Auditability
Trust in privacy tools requires transparency. The core logic of uViewDoc relies on battle-tested open-source libraries like Mozilla's PDF.js. No obfuscated binaries.
View Source on GitHubGDPR Exempt
Because we collect zero user data and process files locally, uViewDoc naturally falls outside GDPR scope.
HIPAA Safe
Medical documents parsed locally never traverse a network, ensuring no PHI is accidentally transmitted.
SOC2 Compliant Default
By operating strictly client-side, the tool inherently complies with zero-trust network mandates.
Ready to inspect?
Stop uploading your sensitive files. Scroll back up and drop a document to begin.
Return to Viewer