pdfmake

pdfmake is a pure-JavaScript PDF generation library that lets applications describe a document — text, tables, columns, images, lists, headers/footers, page breaks — as a plain JSON document-definition object, then renders it to a PDF either on the server (Node.js) or directly in the browser, using the same document-definition format and rendering pipeline in both environments. Built on top of pdfkit for low-level PDF primitives, it adds a full layout engine handling line-wrapping, auto/fixed/star-sized table columns, row/col-spans, repeating table headers across page breaks, snaking (newspaper-style) columns, and style inheritance.

It ships helper methods for opening, downloading, or printing a generated PDF directly from browser code, along with font-embedding support and a hosted playground for experimenting with document definitions interactively. With over 12,000 GitHub stars and 2M+ weekly npm downloads, it’s one of the most widely used JavaScript PDF-generation libraries for invoices, reports, and other structured documents.

What You Get

• A declarative JSON document-definition format describing text, tables, columns, images, and layout
• A full table layout engine: auto/fixed/star-sized widths, col/row-spans, repeating headers, snaking columns
• Identical rendering pipeline usable server-side (Node.js) or client-side (browser) from the same document definition
• Font embedding and custom font support beyond the bundled standard fonts
• Browser helper methods (open, download, print) for delivering generated PDFs directly to users

Common Use Cases

• Generating invoices, receipts, and financial reports server-side from application data
• Letting end users export a report or document to PDF directly in the browser without a server round-trip
• Producing multi-page documents with repeating table headers, page numbers, and dynamic headers/footers
• Building printable, structured business documents (contracts, certificates, tickets) from templated JSON definitions

Under The Hood

Architecture - src/index.js wires together a pipeline of dedicated classes: DocPreprocessor normalizes the raw document-definition object, LayoutBuilder walks it to compute measured lines/pages via TextInlines, TableProcessor/columnCalculator, and DocumentContext, and ElementWriter/Renderer emit the final drawing calls onto a pdfkit document; OutputDocument/OutputDocumentServer split server-only output paths (streams, buffers) from the shared browser/server core, and virtual-fs.js provides an in-memory filesystem abstraction so the same font/asset-loading code works in both environments.

Tech Stack - Pure JavaScript (~8,000 lines across src/) built on pdfkit for low-level PDF primitives plus linebreak for Unicode-aware line breaking and xmldoc for parsing embedded SVG/XML content; Babel compiles the ESM source to both a Node.js CommonJS build (js/) and a bundled browser build (build/pdfmake.js) from the same src/ tree, which is what enables the isomorphic API.

Code Quality - tests/unit (20 files) covers layout primitives (line-wrapping, table/column calculation, style inheritance) with Mocha, tests/integration compares rendered PDFs against fixtures, and tests/browser exercises the browser bundle specifically; ESLint with eslint-plugin-jsdoc enforces documented public APIs across the large layout-engine codebase.

API Design - The single-entry-point createPdf(docDefinition) call and declarative JSON document format keep the common case simple (pass an object, get a PDF), and the same definitions work unchanged between Node and browser; the tradeoff is that achieving precise layouts (nested tables, multi-column flows, custom fonts) requires learning pdfmake’s own JSON schema rather than an imperative drawing API, which raises the ceiling but not the floor of the learning curve.