Bulk image watermark with NodeJS and Jimp

It's been almost two months since I released Mama Culinar - my wife's culinary blog which is based on WordPress. Yesterday I got a new challenging request from her - she wanted to have a watermark on the images attached to recipes.

Now you will say that there are many plugins out there which help you do exactly this but when I checked, most of them were outdated - they weren't tested with the past three versions of WordPress or had too negative reviews.

So I decided to use another approach - why not code it myself? I chose NodeJS as I am familiar with its APIs and it looks like the best options when it comes to heavy data manipulation due to its asynchronous nature.

I started by creating a new NodeJS project:

mkdir bulk-watermark && cd bulk-watermark && yarn init -y

The command above creates a new folder called "bulk-watermark", changes the current working directory to this new folder and creates a "package.json" file with default parameters. You can and will edit this file later.

Then I installed the dependencies. I needed jimp, glob and dotenv so I installed these packages:

yarn add jimp glob dotenv

Then I created a new .env file which contains my environmental variables (or configuration settings):

# Turns on/off console logging

# Absolute path to the file which will be used as a watermark

# Absolute path to the folder containing the images

# Spacing between the watermark and the image edges
# in percentage (from 0 to 100)

# Opacity of the watermark when placed over the image

# Ratio in percentage (from 0 to 100) relative to
# the watermarked image which is used to resize
# the watermark if the watermark is larger

# The position of the watermark.
# Possible values are:
# N (north),
# NE (northeast),
# E (east),
# SE (southeast),
# S (south),
# SW (southwest),
# W (west),
# NW (northwest),
# C (center)

# A pipe split words which shouldn't be found in the images file names

Then I created a new index.js file which contains the actual script:

#!/usr/bin/env node

const { resolve } = require('path');

const jimp = require('jimp');
const glob = require('glob');
const dotnev = require('dotenv');


const {
} = process.env;

 * Read an environmental variable,
 * convert it to a number
 * and then limit it between
 * a min and a max values
 * @param {string} variable
 * @param {number} defaultValue
 * @param {number} min
 * @param {number} max
 * @returns {number}
const useEnvNumber = (variable, defaultValue, min, max) => {
	if (typeof variable === 'undefined' || variable === '') {
		return defaultValue;

	const num = Number(variable);

	return num > max ? max : num < min ? min : num;

const ratio = useEnvNumber(WATERMARK_RESIZE_RATIO, 10, 0, 100);
const spacing = useEnvNumber(WATERMARK_SPACING, 2, 0, 100);
const opacity = useEnvNumber(WATERMARK_OPACITY, 0.9, 0, 1);
const shouldDebug = WATERMARK_DEBUG === 'true';

	shouldDebug && console.error('No images found.');

	shouldDebug && console.error('No watermark found.');

const IMAGES = glob.sync(`${resolve(WATERMARK_IMAGES)}/**/*.{jpeg,jpg,png}`).filter(filename => {
	const name = filename.toLowerCase();

	if (typeof WATERMARK_IGNORE === 'undefined' || WATERMARK_IGNORE === '') {
		return true;

	const words = WATERMARK_IGNORE.split('|');
	const matches = words.filter(word => name.includes(word));

	return matches.length === 0;

 * Get the x and y coordinates of the watermark
 * @param {jimp} image
 * @param {jimp} watermark
 * @param {number} xMargin
 * @param {number} yMargin
 * @returns {Record<'x' | 'y', number>}
const getCoords = (image, watermark, xMargin, yMargin) => {
	const imageWidth = image.bitmap.width;
	const imageHeight = image.bitmap.height;
	const watermarkWidth = watermark.bitmap.width;
	const watermarkHeight = watermark.bitmap.height;

		case 'N':
			return {
				x: imageWidth / 2 - watermarkWidth / 2,
				y: yMargin
		case 'NE':
			return {
				x: imageWidth - watermarkWidth - xMargin,
				y: yMargin
		case 'E':
			return {
				x: imageWidth - watermarkWidth - xMargin,
				y: imageHeight / 2 - watermarkHeight / 2
		case 'SE':
			return {
				x: imageWidth - watermarkWidth - xMargin,
				y: imageHeight - watermarkHeight - yMargin
		case 'S':
			return {
				x: imageWidth / 2 - watermarkWidth / 2,
				y: imageHeight - watermarkHeight - yMargin
		case 'SW':
			return {
				x: xMargin,
				y: imageHeight - watermarkHeight - yMargin
		case 'W':
			return {
				x: xMargin,
				y: imageHeight / 2 - watermarkHeight / 2
		case 'NW':
			return {
				x: xMargin,
				y: yMargin
		case 'C':
			return {
				x: imageWidth / 2 - watermarkWidth / 2,
				y: imageHeight / 2 - watermarkHeight / 2
			return {
				x: 0,
				y: 0

 * Place a watermark over an image
 * @param {string} src
async function waterMark(src) {
	shouldDebug && console.log('Start processing ' + src);

	const image = await jimp.read(src).catch(() => null);
	const watermark = await jimp.read('./icon.png').catch(() => null);

	if (!image || !watermark) {

	watermark.resize(image.bitmap.width / ratio, jimp.AUTO);

	const xMargin = (image.bitmap.width * spacing) / 100;
	const yMargin = (image.bitmap.width * spacing) / 100;

	const { x, y } = getCoords(image, watermark, xMargin, yMargin);

	image.composite(watermark, x, y, {
		mode: jimp.BLEND_SOURCE_OVER,
		opacityDest: 1,
		opacitySource: opacity

	await image.writeAsync(src).catch(e => {
		shouldDebug && console.error(e);

	shouldDebug && console.log('Done processing ' + src);
	shouldDebug && console.log('---');

(async () => {
	for (const image of IMAGES) {
		await waterMark(image);


Let's explain what this does:

First I import the resolve method from the NodeJS' built-in fs module. Then I import the 3rd party modules which were previously installed - jimp, glob and dotenv. Then I configure my environment by calling the dotenv.config() method. This reads the .env file which I created and allows me to read the variables using the NodeJS' process env object.

Then I define the useEnvNumber function which essentially reads an environment variable defined in the .env file and returns it's number representation. The function also uses a default value if the environmental variable is missing and limits the value of the variable between a min and max values.

Then I use the useEnvNumber function to define some variables which will be used in the image manipulations later.

If the WATERMARK_IMAGES and WATERMARK_IMAGE variables are missing, I stop the execution and print an error. Without those two variables defined the script can not continue.

Then I define the IMAGES variable which is an array which contains all images located in the WATERMARK_IMAGES folder. This function takes care of extracting only image files and then filters them based on the WATERMARK_IGNORE words.

The I define the waterMark asynchronous function which accepts a path to an image as an argument and does all the magic:

  1. First it passes the image and the watermark to jimp - the library returns Jimp objects which will be manipulated and used later on.
  2. Then the watermark image is resized based on the size of the large image and the WATERMARK_RESIZE_RATIO variable.
  3. Then I define the X and Y margins of the watermark.
  4. Then based on the WATERMARK_POSITION variable I get the x and y coordinates of the position of the watermark over the image.
  5. Then I use jimp's composite method to place the watermark over the image.
  6. At the end I call the waterMark function for all of the images in the provided WATERMARK_IMAGES folder.

In my case I had a 1.86 GB worth of images with various sizes and all of them needed a watermark.

Using a late 2015 MB Pro with i7 and 16 GB of RAM, this task took exactly 26.5 minutes to complete but you know at least my wife is happy!

Two notes:

  1. The script does not support files with special characters in their names (or files with name in language different than English, for example Bulgarian).
  2. The script overwrites the input images so you probably need to back your images up before starting the watermarking process.

Go back

Send me your message

Trusted by

  • Duke University brand image
  • Emailio brand image
  • E.ON brand image
  • Kinetik Automotive brand image
  • Robert Ladkani brand image
  • SOD 64 brand image
  • Three11 brand image
  • dmarcian brand image
  • htmlBurger brand image
  • htmlBoutique brand image
  • 2create brand image