PDF.js

PDF.js is a community-driven, Mozilla-supported JavaScript library that parses and renders Portable Document Format (PDF) files entirely with web standards, requiring no native plugins. Distributed on npm as pdfjs-dist, it powers Firefox’s built-in PDF viewer and countless web applications that need to display, inspect, or extract content from PDFs.

The library exposes a document API for loading PDFs and rendering their pages to an HTML5 canvas, along with text-layer and annotation-layer building blocks for selectable text, links, and interactive form fields. Heavy parsing runs in a Web Worker so the main thread stays responsive, and the package ships both modern and legacy builds to support a wide range of browsers.

What You Get

• A prebuilt distribution (pdfjs-dist) with pdf.js and a separate pdf.worker for off-main-thread parsing
• A promise-based document/page API (getDocument, page.render) for loading and rendering PDFs to canvas
• Text-layer and annotation-layer helpers for selectable text, hyperlinks, and interactive form fields
• Modern and legacy build variants covering both current and older browser targets
• A complete, ready-to-embed generic web viewer plus a large collection of usage examples

Common Use Cases

• Embedding an in-browser PDF viewer without relying on native browser plugins
• Rendering PDF pages to canvas or images for previews and thumbnails
• Extracting text content from PDFs for search, indexing, or accessibility
• Displaying and filling interactive PDF forms and annotations in a web app

Under The Hood

Architecture — PDF.js is split across three source layers under src/: core/ parses the raw PDF byte stream (objects, xref tables, fonts, CMaps, content-stream operators, image and stream decoders in files like document.js, parser.js, and crypto.js), display/ turns parsed operator lists into rendered output via api.js (the public getDocument/PDFDocumentProxy/PDFPageProxy surface) and canvas.js (the 2D canvas renderer), and shared/ holds constants and utilities used by both. The expensive parsing work runs inside a Web Worker (pdf.worker.js), communicating with the display layer over a structured message transport so the UI thread only receives operator lists to paint. Text and annotation layers (text_layer.js, annotation_layer.js) are rendered as DOM overlays aligned to the canvas.

Tech Stack — The library is written in modern JavaScript (ES modules, "type": "module") with no runtime dependencies in the published package. The build pipeline uses Gulp, Webpack, and Babel (with core-js polyfills for the legacy build) to bundle src/ into the production pdf.js/pdf.worker.js scripts; TypeScript declaration files are generated via tsc for typed consumers. Node.js 22.13+ or 24+ is required for development. Localization uses Fluent, and native canvas rendering for Node tests goes through @napi-rs/canvas.

Code Quality — The repository has an extensive test suite: 71 unit-test modules under test/unit/, plus Puppeteer-driven integration tests, reference (pixel-comparison) tests, and font tests, all wired into GitHub Actions with Codecov coverage tracking across three workflows. Code is linted with ESLint (including no-unsanitized and regexp plugins), formatted with Prettier, and style-checked with Stylelint. With over 23,000 commits and 500+ contributors over more than a decade, the codebase is mature and rigorously maintained.

API Design — The public API is deliberately small and promise-based: getDocument(src) returns a loading task resolving to a document proxy, doc.getPage(n) returns a page proxy, and page.render({ canvasContext, viewport }) paints it. Worker setup is handled by assigning GlobalWorkerOptions.workerSrc, the one piece of required boilerplate. The API is thoroughly documented at mozilla.github.io/pdf.js/api and backed by interactive examples and an examples folder covering components, image decoders, Node usage, and webpack integration, which flattens the otherwise steep learning curve of working with a format as complex as PDF.