pdf-raster

Introduction

pdf-raster renders PDF pages to PNG, JPEG or WebP buffers in Node.js and Bun.

pdf-raster turns PDF pages into encoded images on the server. It runs PDFium, the PDF engine from Chromium, inside a native addon written in Rust with napi-rs. You call one function, convert(), and get back one image buffer per page.

import { convert } from "pdf-raster";

const pages = await convert("./report.pdf");
// pages[0].data is a PNG Buffer of the first page
  1. 01

    Input

    path, Buffer, Uint8Array or ArrayBuffer

  2. 02locked

    Render

    PDFium, one page at a time behind a process-wide lock

  3. 03parallel

    Encode

    PNG, JPEG or WebP on up to 4 threads per call

  4. 04

    Result

    ConvertedPage[] with bytes, size and DPI

When to use it

Use pdf-raster when the next step in your system needs page images instead of PDF bytes. For example:

  • OCR engines such as Tesseract.js
  • vision models that accept image input
  • page previews and thumbnails in a web app
  • batch jobs that store or index documents as images

What it does not do

  • It does not run in browsers, Edge runtimes or WebAssembly. It needs Node.js 24 or newer, or Bun.
  • It does not edit PDFs, fill forms or extract text.

Output

Each page comes back as a ConvertedPage:

type ConvertedPage = {
  pageIndex: number; // zero-based index in the source PDF
  data: Buffer; // the encoded image
  mimeType: "image/png" | "image/jpeg" | "image/webp";
  width: number; // pixels
  height: number; // pixels
  dpi: number;
};

The extra fields describe data, so you can store, serve or forward the bytes without decoding the image first.

Next steps

On this page