I do this npm i -S json-size and then somewhere in my code.

import jsonSize from 'json-size'
console.log(jsonSize({foo: 'bar'})) //=> 13

This is what json-size is all about. github.com/bendrucker/json-size/blob/master/index…

'use strict'

var bytes = require('utf8-length')

module.exports = function jsonSize (value) {
  return bytes(JSON.stringify(value))
}

json-size has a dependency of utf8-lenghth which is nothing more than this. github.com/substack/utf8-length/blob/master/index…

module.exports = function(s) {
  return ~-encodeURI(s).split(/%..|./).length
}

We can do it all by ourself now.

const bytes = (s) => {
  return ~-encodeURI(s).split(/%..|./).length
}
const jsonSize = (s) => {​
  return bytes(JSON.stringify(s))
}
console.log(jsonSize({foo: 'bar'})) //=> 13

What's this tilde operator? joezimjs.com/javascript/great-mystery-of-the-tilde

MongoDB uses BSON to store objects so the calculated size will need to depend on BSON. Take a look at https://www.npmjs.com/package/bson:

import bson from 'bson'
const BSON = new bson.BSONPure.BSON()

BSON.calculateObjectSize({foo: 'bar'}) // => 18 bytes

This value will be close but still approximate since Mongo uses further optimizations.

If you are about to exceed the MongoDB document size limit, you sould rethink your data structures.