ZebraCrossing

ZebraCrossing is a wrapper around the zxing CLI. It supports all sensible options of the zxing CLI and lets you pass buffers as well as file paths (but see the note on this).

It's still not fast, it still a bit hacky, it's still your best bet for reading barcodes.

const ZebraCrossing = require('zebra-crossing');
ZebraCrossing.read(fs.readFileSync('image.png'), { pureBarcode: true })
	.then(data => console.log(data));

Example output:

{
	raw: Buffer.from('Hello world!'),
	parsed: Buffer.from('Hello world!'),
	format: 'QR_CODE',
	type: 'TEXT',
	points: [ [774.0,461.0], [774.0,405.0], [830.0,405.0] ]
}

Requirements

The dirty work of actually parsing the barcodes is done by zxing, a Java library, so you'll need to have a Java Runtime Environment installed.

API

ZebraCrossing only has one method: read(file<String|Buffer>, [options]).

If file is a string, it will be assumed to be a file path that will be passed to zxing. If file is a buffer, its contents will be written to a temporary file whose path will be passed to zxing.

options can be an object like this:

{
  // Use the TRY_HARDER hint
  tryHarder: true,
  // Image is a direct monochrome image of a barcode, not a photo
  pureBarcode: true,
  // Only read retail barcodes (EAN and UPC)
  productsOnly: true,
  // Find multiple barcodes in the same image
  multi: true,
  // Crop by the specified pixel values before searching for barcodes
  crop: { left: 10, top: 20, width: 30, height: 40 },
  // Only search for the specified barcode formats
  // See ZebraCrossing.FORMAT for supported formats
  possibleFormats: [ 'AZTEC', 'QR_CODE' ],
}

ZebraCrossing.read() returns a Promise that resolves with a { raw, parsed, format, type, points } object if a barcode was found or null if none was found. If multi: true was set in the options, it resolves with an array of result objects, or the empty array if no barcodes were found.

A note on file paths

Because zxing's CLI only accepts file paths, buffers are written to a temporary file on disk before they're parsed. This means that you should always pass a file path if you can. Accepting buffers is just for your convenience.