How do I create a GUID / UUID?

[Edited 2023-03-05 to reflect latest best-practices for producing RFC4122-compliant UUIDs]

crypto.randomUUID() is now standard on all modern browsers and JS runtimes. However, because new browser APIs are restricted to secure contexts, this method is only available to pages served locally (localhost or 127.0.0.1) or over HTTPS.

For readers interested in other UUID versions, generating UUIDs on legacy platforms or in non-secure contexts, there is the uuid module. It is well-tested and supported.

Failing the above, there is this method (based on the original answer to this question):

function uuidv4() {
  return "10000000-1000-4000-8000-100000000000".replace(/[018]/g, c =>
    (c ^ crypto.getRandomValues(new Uint8Array(1))[0] & 15 >> c / 4).toString(16)
  );
}

console.log(uuidv4());

Note: The use of any UUID generator that relies on Math.random() is strongly discouraged (including snippets featured in previous versions of this answer) for reasons best explained here. TL;DR: solutions based on Math.random() do not provide good uniqueness guarantees.

UUIDs (Universally Unique IDentifier), also known as GUIDs (Globally Unique IDentifier), according to RFC 4122, are identifiers designed to provide certain uniqueness guarantees.

While it is possible to implement RFC-compliant UUIDs in a few lines of JavaScript code (e.g., see @broofa's answer, below) there are several common pitfalls:

• Invalid id format (UUIDs must be of the form "xxxxxxxx-xxxx-Mxxx-Nxxx-xxxxxxxxxxxx", where x is one of [0-9, a-f] M is one of [1-5], and N is [8, 9, a, or b]
• Use of a low-quality source of randomness (such as Math.random)

Thus, developers writing code for production environments are encouraged to use a rigorous, well-maintained implementation such as the uuid module.

I really like how clean Broofa's answer is, but it's unfortunate that poor implementations of Math.random leave the chance for collision.

Here's a similar RFC4122 version 4 compliant solution that solves that issue by offsetting the first 13 hex numbers by a hex portion of the timestamp, and once depleted offsets by a hex portion of the microseconds since pageload. That way, even if Math.random is on the same seed, both clients would have to generate the UUID the exact same number of microseconds since pageload (if high-perfomance time is supported) AND at the exact same millisecond (or 10,000+ years later) to get the same UUID:

function generateUUID() { // Public Domain/MIT
    var d = new Date().getTime();//Timestamp
    var d2 = ((typeof performance !== 'undefined') && performance.now && (performance.now()*1000)) || 0;//Time in microseconds since page-load or 0 if unsupported
    return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
        var r = Math.random() * 16;//random number between 0 and 16
        if(d > 0){//Use timestamp until depleted
            r = (d + r)%16 | 0;
            d = Math.floor(d/16);
        } else {//Use microseconds since page-load if supported
            r = (d2 + r)%16 | 0;
            d2 = Math.floor(d2/16);
        }
        return (c === 'x' ? r : (r & 0x3 | 0x8)).toString(16);
    });
}

var onClick = function(){
    document.getElementById('uuid').textContent = generateUUID();
}
onClick();

#uuid { font-family: monospace; font-size: 1.5em; }

<p id="uuid"></p>
<button id="generateUUID" onclick="onClick();">Generate UUID</button>

Here's a fiddle to test.

Modernized snippet for ES6

const generateUUID = () => {
  let
    d = new Date().getTime(),
    d2 = ((typeof performance !== 'undefined') && performance.now && (performance.now() * 1000)) || 0;
  return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, c => {
    let r = Math.random() * 16;
    if (d > 0) {
      r = (d + r) % 16 | 0;
      d = Math.floor(d / 16);
    } else {
      r = (d2 + r) % 16 | 0;
      d2 = Math.floor(d2 / 16);
    }
    return (c == 'x' ? r : (r & 0x7 | 0x8)).toString(16);
  });
};

const onClick = (e) => document.getElementById('uuid').textContent = generateUUID();

document.getElementById('generateUUID').addEventListener('click', onClick);

onClick();

#uuid { font-family: monospace; font-size: 1.5em; }

<p id="uuid"></p>
<button id="generateUUID">Generate UUID</button>

broofa's answer is pretty slick, indeed - impressively clever, really... RFC4122 compliant, somewhat readable, and compact. Awesome!

But if you're looking at that regular expression, those many replace() callbacks, toString()'s and Math.random() function calls (where he's only using four bits of the result and wasting the rest), you may start to wonder about performance. Indeed, joelpt even decided to toss out an RFC for generic GUID speed with generateQuickGUID.

But, can we get speed and RFC compliance? I say, YES! Can we maintain readability? Well... Not really, but it's easy if you follow along.

But first, my results, compared to broofa, guid (the accepted answer), and the non-rfc-compliant generateQuickGuid:

                  Desktop   Android
           broofa: 1617ms   12869ms
               e1:  636ms    5778ms
               e2:  606ms    4754ms
               e3:  364ms    3003ms
               e4:  329ms    2015ms
               e5:  147ms    1156ms
               e6:  146ms    1035ms
               e7:  105ms     726ms
             guid:  962ms   10762ms
generateQuickGuid:  292ms    2961ms
  - Note: 500k iterations, results will vary by browser/CPU.

So by my 6th iteration of optimizations, I beat the most popular answer by over 12 times, the accepted answer by over 9 times, and the fast-non-compliant answer by 2-3 times. And I'm still RFC 4122 compliant.

Interested in how? I've put the full source on http://jsfiddle.net/jcward/7hyaC/3/ and on https://jsben.ch/xczxS

For an explanation, let's start with broofa's code:

function broofa() {
    return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
        var r = Math.random()*16|0, v = c == 'x' ? r : (r&0x3|0x8);
        return v.toString(16);
    });
}

console.log(broofa())

So it replaces x with any random hexadecimal digit, y with random data (except forcing the top two bits to 10 per the RFC spec), and the regex doesn't match the - or 4 characters, so he doesn't have to deal with them. Very, very slick.

The first thing to know is that function calls are expensive, as are regular expressions (though he only uses 1, it has 32 callbacks, one for each match, and in each of the 32 callbacks it calls Math.random() and v.toString(16)).

The first step toward performance is to eliminate the RegEx and its callback functions and use a simple loop instead. This means we have to deal with the - and 4 characters whereas broofa did not. Also, note that we can use String Array indexing to keep his slick String template architecture:

function e1() {
    var u='',i=0;
    while(i++<36) {
        var c='xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'[i-1],r=Math.random()*16|0,v=c=='x'?r:(r&0x3|0x8);
        u+=(c=='-'||c=='4')?c:v.toString(16)
    }
    return u;
}

console.log(e1())

Basically, the same inner logic, except we check for - or 4, and using a while loop (instead of replace() callbacks) gets us an almost 3X improvement!

The next step is a small one on the desktop but makes a decent difference on mobile. Let's make fewer Math.random() calls and utilize all those random bits instead of throwing 87% of them away with a random buffer that gets shifted out each iteration. Let's also move that template definition out of the loop, just in case it helps:

function e2() {
    var u='',m='xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx',i=0,rb=Math.random()*0xffffffff|0;
    while(i++<36) {
        var c=m[i-1],r=rb&0xf,v=c=='x'?r:(r&0x3|0x8);
        u+=(c=='-'||c=='4')?c:v.toString(16);rb=i%8==0?Math.random()*0xffffffff|0:rb>>4
    }
    return u
}

console.log(e2())

This saves us 10-30% depending on platform. Not bad. But the next big step gets rid of the toString function calls altogether with an optimization classic - the look-up table. A simple 16-element lookup table will perform the job of toString(16) in much less time:

function e3() {
    var h='0123456789abcdef';
    var k='xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx';
    /* same as e4() below */
}
function e4() {
    var h=['0','1','2','3','4','5','6','7','8','9','a','b','c','d','e','f'];
    var k=['x','x','x','x','x','x','x','x','-','x','x','x','x','-','4','x','x','x','-','y','x','x','x','-','x','x','x','x','x','x','x','x','x','x','x','x'];
    var u='',i=0,rb=Math.random()*0xffffffff|0;
    while(i++<36) {
        var c=k[i-1],r=rb&0xf,v=c=='x'?r:(r&0x3|0x8);
        u+=(c=='-'||c=='4')?c:h[v];rb=i%8==0?Math.random()*0xffffffff|0:rb>>4
    }
    return u
}

console.log(e4())

The next optimization is another classic. Since we're only handling four bits of output in each loop iteration, let's cut the number of loops in half and process eight bits in each iteration. This is tricky since we still have to handle the RFC compliant bit positions, but it's not too hard. We then have to make a larger lookup table (16x16, or 256) to store 0x00 - 0xFF, and we build it only once, outside the e5() function.

var lut = []; for (var i=0; i<256; i++) { lut[i] = (i<16?'0':'')+(i).toString(16); }
function e5() {
    var k=['x','x','x','x','-','x','x','-','4','x','-','y','x','-','x','x','x','x','x','x'];
    var u='',i=0,rb=Math.random()*0xffffffff|0;
    while(i++<20) {
        var c=k[i-1],r=rb&0xff,v=c=='x'?r:(c=='y'?(r&0x3f|0x80):(r&0xf|0x40));
        u+=(c=='-')?c:lut[v];rb=i%4==0?Math.random()*0xffffffff|0:rb>>8
    }
    return u
}

console.log(e5())

I tried an e6() that processes 16-bits at a time, still using the 256-element LUT, and it showed the diminishing returns of optimization. Though it had fewer iterations, the inner logic was complicated by the increased processing, and it performed the same on desktop, and only ~10% faster on mobile.

The final optimization technique to apply - unroll the loop. Since we're looping a fixed number of times, we can technically write this all out by hand. I tried this once with a single random variable, r, that I kept reassigning, and performance tanked. But with four variables assigned random data up front, then using the lookup table, and applying the proper RFC bits, this version smokes them all:

var lut = []; for (var i=0; i<256; i++) { lut[i] = (i<16?'0':'')+(i).toString(16); }
function e7()
{
    var d0 = Math.random()*0xffffffff|0;
    var d1 = Math.random()*0xffffffff|0;
    var d2 = Math.random()*0xffffffff|0;
    var d3 = Math.random()*0xffffffff|0;
    return lut[d0&0xff]+lut[d0>>8&0xff]+lut[d0>>16&0xff]+lut[d0>>24&0xff]+'-'+
    lut[d1&0xff]+lut[d1>>8&0xff]+'-'+lut[d1>>16&0x0f|0x40]+lut[d1>>24&0xff]+'-'+
    lut[d2&0x3f|0x80]+lut[d2>>8&0xff]+'-'+lut[d2>>16&0xff]+lut[d2>>24&0xff]+
    lut[d3&0xff]+lut[d3>>8&0xff]+lut[d3>>16&0xff]+lut[d3>>24&0xff];
}

console.log(e7())

Modualized: http://jcward.com/UUID.js - UUID.generate()

The funny thing is, generating 16 bytes of random data is the easy part. The whole trick is expressing it in string format with RFC compliance, and it's most tightly accomplished with 16 bytes of random data, an unrolled loop and lookup table.

I hope my logic is correct -- it's very easy to make a mistake in this kind of tedious bit work. But the outputs look good to me. I hope you enjoyed this mad ride through code optimization!

Be advised: my primary goal was to show and teach potential optimization strategies. Other answers cover important topics such as collisions and truly random numbers, which are important for generating good UUIDs.

Use:

let uniqueId = Date.now().toString(36) + Math.random().toString(36).substring(2);

document.getElementById("unique").innerHTML =
  Math.random().toString(36).substring(2) + (new Date()).getTime().toString(36);

<div id="unique">
</div>

If IDs are generated more than 1 millisecond apart, they are 100% unique.

If two IDs are generated at shorter intervals, and assuming that the random method is truly random, this would generate IDs that are 99.99999999999999% likely to be globally unique (collision in 1 of 10^15).

You can increase this number by adding more digits, but to generate 100% unique IDs you will need to use a global counter.

If you need RFC compatibility, this formatting will pass as a valid version 4 GUID:

let u = Date.now().toString(16) + Math.random().toString(16) + '0'.repeat(16);
let guid = [u.substr(0,8), u.substr(8,4), '4000-8' + u.substr(13,3), u.substr(16,12)].join('-');

let u = Date.now().toString(16)+Math.random().toString(16)+'0'.repeat(16);
let guid = [u.substr(0,8), u.substr(8,4), '4000-8' + u.substr(13,3), u.substr(16,12)].join('-');
document.getElementById("unique").innerHTML = guid;

<div id="unique">
</div>

The above code follow the intention, but not the letter of the RFC. Among other discrepancies it's a few random digits short. (Add more random digits if you need it) The upside is that this is really fast :) You can test validity of your GUID here

Here's some code based on RFC 4122, section 4.4 (Algorithms for Creating a UUID from Truly Random or Pseudo-Random Number).

function createUUID() {
    // http://www.ietf.org/rfc/rfc4122.txt
    var s = [];
    var hexDigits = "0123456789abcdef";
    for (var i = 0; i < 36; i++) {
        s[i] = hexDigits.substr(Math.floor(Math.random() * 0x10), 1);
    }
    s[14] = "4";  // bits 12-15 of the time_hi_and_version field to 0010
    s[19] = hexDigits.substr((s[19] & 0x3) | 0x8, 1);  // bits 6-7 of the clock_seq_hi_and_reserved to 01
    s[8] = s[13] = s[18] = s[23] = "-";

    var uuid = s.join("");
    return uuid;
}

This is the fastest GUID-like string generator method in the format XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX. It does not generate a standard-compliant GUID.

Ten million executions of this implementation take just 32.5 seconds, which is the fastest I've ever seen in a browser (the only solution without loops/iterations).

The function is as simple as:

/**
 * Generates a GUID string.
 * @returns {string} The generated GUID.
 * @example af8a8416-6e18-a307-bd9c-f2c947bbb3aa
 * @author Slavik Meltser.
 * @link http://slavik.meltser.info/?p=142
 */
function guid() {
    function _p8(s) {
        var p = (Math.random().toString(16)+"000000000").substr(2,8);
        return s ? "-" + p.substr(0,4) + "-" + p.substr(4,4) : p ;
    }
    return _p8() + _p8(true) + _p8(true) + _p8();
}

To test the performance, you can run this code:

console.time('t');
for (var i = 0; i < 10000000; i++) {
    guid();
};
console.timeEnd('t');

I'm sure most of you will understand what I did there, but maybe there is at least one person that will need an explanation:

The algorithm:

• The Math.random() function returns a decimal number between 0 and 1 with 16 digits after the decimal fraction point (for example 0.4363923368509859).
• Then we take this number and convert it to a string with base 16 (from the example above we'll get 0.6fb7687f). Math.random().toString(16).
• Then we cut off the 0. prefix (0.6fb7687f => 6fb7687f) and get a string with eight hexadecimal characters long. (Math.random().toString(16).substr(2,8).
• Sometimes the Math.random() function will return shorter number (for example 0.4363), due to zeros at the end (from the example above, actually the number is 0.4363000000000000). That's why I'm appending to this string "000000000" (a string with nine zeros) and then cutting it off with substr() function to make it nine characters exactly (filling zeros to the right).
• The reason for adding exactly nine zeros is because of the worse case scenario, which is when the Math.random() function will return exactly 0 or 1 (probability of 1/10^16 for each one of them). That's why we needed to add nine zeros to it ("0"+"000000000" or "1"+"000000000"), and then cutting it off from the second index (third character) with a length of eight characters. For the rest of the cases, the addition of zeros will not harm the result because it is cutting it off anyway. Math.random().toString(16)+"000000000").substr(2,8).

The assembly:

• The GUID is in the following format XXXXXXXX-XXXX-XXXX-XXXX-XXXXXXXXXXXX.
• I divided the GUID into four pieces, each piece divided into two types (or formats): XXXXXXXX and -XXXX-XXXX.
• Now I'm building the GUID using these two types to assemble the GUID with call four pieces, as follows: XXXXXXXX -XXXX-XXXX -XXXX-XXXX XXXXXXXX.
• To differ between these two types, I added a flag parameter to a pair creator function _p8(s), the s parameter tells the function whether to add dashes or not.
• Eventually we build the GUID with the following chaining: _p8() + _p8(true) + _p8(true) + _p8(), and return it.

Link to this post on my blog

Enjoy! :-)

Here is a totally non-compliant but very performant implementation to generate an ASCII-safe GUID-like unique identifier.

function generateQuickGuid() {
    return Math.random().toString(36).substring(2, 15) +
        Math.random().toString(36).substring(2, 15);
}

Generates 26 [a-z0-9] characters, yielding a UID that is both shorter and more unique than RFC compliant GUIDs. Dashes can be trivially added if human-readability matters.

Here are usage examples and timings for this function and several of this question's other answers. The timing was performed under Chrome m25, 10 million iterations each.

>>> generateQuickGuid()
"nvcjf1hs7tf8yyk4lmlijqkuo9"
"yq6gipxqta4kui8z05tgh9qeel"
"36dh5sec7zdj90sk2rx7pjswi2"
runtime: 32.5s

>>> GUID() // John Millikin
"7a342ca2-e79f-528e-6302-8f901b0b6888"
runtime: 57.8s

>>> regexGuid() // broofa
"396e0c46-09e4-4b19-97db-bd423774a4b3"
runtime: 91.2s

>>> createUUID() // Kevin Hakanson
"403aa1ab-9f70-44ec-bc08-5d5ac56bd8a5"
runtime: 65.9s

>>> UUIDv4() // Jed Schmidt
"f4d7d31f-fa83-431a-b30c-3e6cc37cc6ee"
runtime: 282.4s

>>> Math.uuid() // broofa
"5BD52F55-E68F-40FC-93C2-90EE069CE545"
runtime: 225.8s

>>> Math.uuidFast() // broofa
"6CB97A68-23A2-473E-B75B-11263781BBE6"
runtime: 92.0s

>>> Math.uuidCompact() // broofa
"3d7b7a06-0a67-4b67-825c-e5c43ff8c1e8"
runtime: 229.0s

>>> bitwiseGUID() // jablko
"baeaa2f-7587-4ff1-af23-eeab3e92"
runtime: 79.6s

>>>> betterWayGUID() // Andrea Turri
"383585b0-9753-498d-99c3-416582e9662c"
runtime: 60.0s

>>>> UUID() // John Fowler
"855f997b-4369-4cdb-b7c9-7142ceaf39e8"
runtime: 62.2s

Here is the timing code.

var r;
console.time('t'); 
for (var i = 0; i < 10000000; i++) { 
    r = FuncToTest(); 
};
console.timeEnd('t');

From sagi shkedy's technical blog:

function generateGuid() {
  var result, i, j;
  result = '';
  for(j=0; j<32; j++) {
    if( j == 8 || j == 12 || j == 16 || j == 20)
      result = result + '-';
    i = Math.floor(Math.random()*16).toString(16).toUpperCase();
    result = result + i;
  }
  return result;
}

There are other methods that involve using an ActiveX control, but stay away from these!

I thought it was worth pointing out that no GUID generator can guarantee unique keys (check the Wikipedia article). There is always a chance of collisions. A GUID simply offers a large enough universe of keys to reduce the change of collisions to almost nil.

Here is a combination of the top voted answer, with a workaround for Chrome's collisions:

generateGUID = (typeof(window.crypto) != 'undefined' &&
                typeof(window.crypto.getRandomValues) != 'undefined') ?
    function() {
        // If we have a cryptographically secure PRNG, use that
        // https://stackoverflow.com/questions/6906916/collisions-when-generating-uuids-in-javascript
        var buf = new Uint16Array(8);
        window.crypto.getRandomValues(buf);
        var S4 = function(num) {
            var ret = num.toString(16);
            while(ret.length < 4){
                ret = "0"+ret;
            }
            return ret;
        };
        return (S4(buf[0])+S4(buf[1])+"-"+S4(buf[2])+"-"+S4(buf[3])+"-"+S4(buf[4])+"-"+S4(buf[5])+S4(buf[6])+S4(buf[7]));
    }

    :

    function() {
        // Otherwise, just use Math.random
        // https://stackoverflow.com/questions/105034/how-to-create-a-guid-uuid-in-javascript/2117523#2117523
        return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
            var r = Math.random()*16|0, v = c == 'x' ? r : (r&0x3|0x8);
            return v.toString(16);
        });
    };

It is on jsbin if you want to test it.

Here's a solution dated Oct. 9, 2011 from a comment by user jed at https://gist.github.com/982883:

UUIDv4 = function b(a){return a?(a^Math.random()*16>>a/4).toString(16):([1e7]+-1e3+-4e3+-8e3+-1e11).replace(/[018]/g,b)}

This accomplishes the same goal as the current highest-rated answer, but in 50+ fewer bytes by exploiting coercion, recursion, and exponential notation. For those curious how it works, here's the annotated form of an older version of the function:

UUIDv4 =

function b(
  a // placeholder
){
  return a // if the placeholder was passed, return
    ? ( // a random number from 0 to 15
      a ^ // unless b is 8,
      Math.random() // in which case
      * 16 // a random number from
      >> a/4 // 8 to 11
      ).toString(16) // in hexadecimal
    : ( // or otherwise a concatenated string:
      [1e7] + // 10000000 +
      -1e3 + // -1000 +
      -4e3 + // -4000 +
      -8e3 + // -80000000 +
      -1e11 // -100000000000,
      ).replace( // replacing
        /[018]/g, // zeroes, ones, and eights with
        b // random hex digits
      )
}

You can use node-uuid. It provides simple, fast generation of RFC4122 UUIDS.

Features:

• Generate RFC4122 version 1 or version 4 UUIDs
• Runs in Node.js and browsers.
• Cryptographically strong random # generation on supporting platforms.
• Small footprint (Want something smaller? Check this out!)

Install Using NPM:

npm install uuid

Or using uuid via a browser:

Download Raw File (uuid v1): https://raw.githubusercontent.com/kelektiv/node-uuid/master/v1.js Download Raw File (uuid v4): https://raw.githubusercontent.com/kelektiv/node-uuid/master/v4.js

Want even smaller? Check this out: https://gist.github.com/jed/982883

Usage:

// Generate a v1 UUID (time-based)
const uuidV1 = require('uuid/v1');
uuidV1(); // -> '6c84fb90-12c4-11e1-840d-7b25c5ee775a'

// Generate a v4 UUID (random)
const uuidV4 = require('uuid/v4');
uuidV4(); // -> '110ec58a-a0f2-4ac4-8393-c866d813b8d1'

// Generate a v5 UUID (namespace)
const uuidV5 = require('uuid/v5');

// ... using predefined DNS namespace (for domain names)
uuidV5('hello.example.com', v5.DNS)); // -> 'fdda765f-fc57-5604-a269-52a7df8164ec'

// ... using predefined URL namespace (for, well, URLs)
uuidV5('http://example.com/hello', v5.URL); // -> '3bbcee75-cecc-5b56-8031-b6641c1ed1f1'

// ... using a custom namespace
const MY_NAMESPACE = '(previously generated unique uuid string)';
uuidV5('hello', MY_NAMESPACE); // -> '90123e1c-7512-523e-bb28-76fab9f2f73d'

ECMAScript 2015 (ES6):

import uuid from 'uuid/v4';
const id = uuid();

One line solution using Blobs.

window.URL.createObjectURL(new Blob([])).substring(31);

The value at the end (31) depends on the length of the URL.

EDIT:

A more compact and universal solution, as suggested by rinogo:

URL.createObjectURL(new Blob([])).slice(-36);

This creates a version 4 UUID (created from pseudo random numbers):

function uuid()
{
   var chars = '0123456789abcdef'.split('');

   var uuid = [], rnd = Math.random, r;
   uuid[8] = uuid[13] = uuid[18] = uuid[23] = '-';
   uuid[14] = '4'; // version 4

   for (var i = 0; i < 36; i++)
   {
      if (!uuid[i])
      {
         r = 0 | rnd()*16;

         uuid[i] = chars[(i == 19) ? (r & 0x3) | 0x8 : r & 0xf];
      }
   }

   return uuid.join('');
}

Here is a sample of the UUIDs generated:

682db637-0f31-4847-9cdf-25ba9613a75c
97d19478-3ab2-4aa1-b8cc-a1c3540f54aa
2eed04c9-2692-456d-a0fd-51012f947136

var uuid = function() {
    var buf = new Uint32Array(4);
    window.crypto.getRandomValues(buf);
    var idx = -1;
    return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c) {
        idx++;
        var r = (buf[idx>>3] >> ((idx%8)*4))&15;
        var v = c == 'x' ? r : (r&0x3|0x8);
        return v.toString(16);
    });
};

This version is based on Briguy37's answer and some bitwise operators to extract nibble sized windows from the buffer.

It should adhere to the RFC Type 4 (random) schema, since I had problems last time parsing non-compliant UUIDs with Java's UUID.

Simple JavaScript module as a combination of best answers in this question.

var crypto = window.crypto || window.msCrypto || null; // IE11 fix

var Guid = Guid || (function() {

  var EMPTY = '00000000-0000-0000-0000-000000000000';

  var _padLeft = function(paddingString, width, replacementChar) {
    return paddingString.length >= width ? paddingString : _padLeft(replacementChar + paddingString, width, replacementChar || ' ');
  };

  var _s4 = function(number) {
    var hexadecimalResult = number.toString(16);
    return _padLeft(hexadecimalResult, 4, '0');
  };

  var _cryptoGuid = function() {
    var buffer = new window.Uint16Array(8);
    crypto.getRandomValues(buffer);
    return [_s4(buffer[0]) + _s4(buffer[1]), _s4(buffer[2]), _s4(buffer[3]), _s4(buffer[4]), _s4(buffer[5]) + _s4(buffer[6]) + _s4(buffer[7])].join('-');
  };

  var _guid = function() {
    var currentDateMilliseconds = new Date().getTime();
    return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(currentChar) {
      var randomChar = (currentDateMilliseconds + Math.random() * 16) % 16 | 0;
      currentDateMilliseconds = Math.floor(currentDateMilliseconds / 16);
      return (currentChar === 'x' ? randomChar : (randomChar & 0x7 | 0x8)).toString(16);
    });
  };

  var create = function() {
    var hasCrypto = crypto != 'undefined' && crypto !== null,
      hasRandomValues = typeof(window.crypto.getRandomValues) != 'undefined';
    return (hasCrypto && hasRandomValues) ? _cryptoGuid() : _guid();
  };

  return {
    newGuid: create,
    empty: EMPTY
  };
})();

// DEMO: Create and show GUID
console.log('1. New Guid:   ' + Guid.newGuid());

// DEMO: Show empty GUID
console.log('2. Empty Guid: ' + Guid.empty);

Usage:

Guid.newGuid()

"c6c2d12f-d76b-5739-e551-07e6de5b0807"

Guid.empty

"00000000-0000-0000-0000-000000000000"

The version below is an adaptation of broofa's answer, but updated to include a "true" random function that uses crypto libraries where available, and the Alea() function as a fallback.

  Math.log2 = Math.log2 || function(n){ return Math.log(n) / Math.log(2); }
  Math.trueRandom = (function() {
  var crypt = window.crypto || window.msCrypto;

  if (crypt && crypt.getRandomValues) {
      // If we have a crypto library, use it
      var random = function(min, max) {
          var rval = 0;
          var range = max - min;
          if (range < 2) {
              return min;
          }

          var bits_needed = Math.ceil(Math.log2(range));
          if (bits_needed > 53) {
            throw new Exception("We cannot generate numbers larger than 53 bits.");
          }
          var bytes_needed = Math.ceil(bits_needed / 8);
          var mask = Math.pow(2, bits_needed) - 1;
          // 7776 -> (2^13 = 8192) -1 == 8191 or 0x00001111 11111111

          // Create byte array and fill with N random numbers
          var byteArray = new Uint8Array(bytes_needed);
          crypt.getRandomValues(byteArray);

          var p = (bytes_needed - 1) * 8;
          for(var i = 0; i < bytes_needed; i++ ) {
              rval += byteArray[i] * Math.pow(2, p);
              p -= 8;
          }

          // Use & to apply the mask and reduce the number of recursive lookups
          rval = rval & mask;

          if (rval >= range) {
              // Integer out of acceptable range
              return random(min, max);
          }
          // Return an integer that falls within the range
          return min + rval;
      }
      return function() {
          var r = random(0, 1000000000) / 1000000000;
          return r;
      };
  } else {
      // From https://web.archive.org/web/20120502223108/http://baagoe.com/en/RandomMusings/javascript/
      // Johannes Baagøe <baagoe@baagoe.com>, 2010
      function Mash() {
          var n = 0xefc8249d;

          var mash = function(data) {
              data = data.toString();
              for (var i = 0; i < data.length; i++) {
                  n += data.charCodeAt(i);
                  var h = 0.02519603282416938 * n;
                  n = h >>> 0;
                  h -= n;
                  h *= n;
                  n = h >>> 0;
                  h -= n;
                  n += h * 0x100000000; // 2^32
              }
              return (n >>> 0) * 2.3283064365386963e-10; // 2^-32
          };

          mash.version = 'Mash 0.9';
          return mash;
      }

      // From http://baagoe.com/en/RandomMusings/javascript/
      function Alea() {
          return (function(args) {
              // Johannes Baagøe <baagoe@baagoe.com>, 2010
              var s0 = 0;
              var s1 = 0;
              var s2 = 0;
              var c = 1;

              if (args.length == 0) {
                  args = [+new Date()];
              }
              var mash = Mash();
              s0 = mash(' ');
              s1 = mash(' ');
              s2 = mash(' ');

              for (var i = 0; i < args.length; i++) {
                  s0 -= mash(args[i]);
                  if (s0 < 0) {
                      s0 += 1;
                  }
                  s1 -= mash(args[i]);
                  if (s1 < 0) {
                      s1 += 1;
                  }
                  s2 -= mash(args[i]);
                  if (s2 < 0) {
                      s2 += 1;
                  }
              }
              mash = null;

              var random = function() {
                  var t = 2091639 * s0 + c * 2.3283064365386963e-10; // 2^-32
                  s0 = s1;
                  s1 = s2;
                  return s2 = t - (c = t | 0);
              };
              random.uint32 = function() {
                  return random() * 0x100000000; // 2^32
              };
              random.fract53 = function() {
                  return random() +
                      (random() * 0x200000 | 0) * 1.1102230246251565e-16; // 2^-53
              };
              random.version = 'Alea 0.9';
              random.args = args;
              return random;

          }(Array.prototype.slice.call(arguments)));
      };
      return Alea();
  }
}());

Math.guid = function() {
    return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(/[xy]/g, function(c)    {
      var r = Math.trueRandom() * 16 | 0,
          v = c == 'x' ? r : (r & 0x3 | 0x8);
      return v.toString(16);
  });
};

JavaScript project on GitHub - https://github.com/LiosK/UUID.js

UUID.js The RFC-compliant UUID generator for JavaScript.

See RFC 4122 http://www.ietf.org/rfc/rfc4122.txt.

Features Generates RFC 4122 compliant UUIDs.

Version 4 UUIDs (UUIDs from random numbers) and version 1 UUIDs (time-based UUIDs) are available.

UUID object allows a variety of access to the UUID including access to the UUID fields.

Low timestamp resolution of JavaScript is compensated by random numbers.

Added in: v15.6.0, v14.17.0 there is a built-in crypto.randomUUID() function.

import * as crypto from "crypto";

const uuid = crypto.randomUUID();

In the browser, crypto.randomUUID() is currently supported in Chromium 92+ and Firefox 95+.

  // RFC 4122
  //
  // A UUID is 128 bits long
  //
  // String representation is five fields of 4, 2, 2, 2, and 6 bytes.
  // Fields represented as lowercase, zero-filled, hexadecimal strings, and
  // are separated by dash characters
  //
  // A version 4 UUID is generated by setting all but six bits to randomly
  // chosen values
  var uuid = [
    Math.random().toString(16).slice(2, 10),
    Math.random().toString(16).slice(2, 6),

    // Set the four most significant bits (bits 12 through 15) of the
    // time_hi_and_version field to the 4-bit version number from Section
    // 4.1.3
    (Math.random() * .0625 /* 0x.1 */ + .25 /* 0x.4 */).toString(16).slice(2, 6),

    // Set the two most significant bits (bits 6 and 7) of the
    // clock_seq_hi_and_reserved to zero and one, respectively
    (Math.random() * .25 /* 0x.4 */ + .5 /* 0x.8 */).toString(16).slice(2, 6),

    Math.random().toString(16).slice(2, 14)].join('-');

For those wanting an RFC 4122 version 4 compliant solution with speed considerations (few calls to Math.random()):

var rand = Math.random;

function UUID() {
    var nbr, randStr = "";
    do {
        randStr += (nbr = rand()).toString(16).substr(3, 6);
    } while (randStr.length < 30);
    return (
        randStr.substr(0, 8) + "-" +
        randStr.substr(8, 4) + "-4" +
        randStr.substr(12, 3) + "-" +
        ((nbr*4|0)+8).toString(16) + // [89ab]
        randStr.substr(15, 3) + "-" +
        randStr.substr(18, 12)
    );
}

console.log( UUID() );

The above function should have a decent balance between speed and randomness.

I wanted to understand broofa's answer, so I expanded it and added comments:

var uuid = function () {
    return 'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx'.replace(
        /[xy]/g,
        function (match) {
            /*
            * Create a random nibble. The two clever bits of this code:
            *
            * - Bitwise operations will truncate floating point numbers
            * - For a bitwise OR of any x, x | 0 = x
            *
            * So:
            *
            * Math.random * 16
            *
            * creates a random floating point number
            * between 0 (inclusive) and 16 (exclusive) and
            *
            * | 0
            *
            * truncates the floating point number into an integer.
            */
            var randomNibble = Math.random() * 16 | 0;

            /*
            * Resolves the variant field. If the variant field (delineated
            * as y in the initial string) is matched, the nibble must
            * match the mask (where x is a do-not-care bit):
            *
            * 10xx
            *
            * This is achieved by performing the following operations in
            * sequence (where x is an intermediate result):
            *
            * - x & 0x3, which is equivalent to x % 3
            * - x | 0x8, which is equivalent to x + 8
            *
            * This results in a nibble between 8 inclusive and 11 exclusive,
            * (or 1000 and 1011 in binary), all of which satisfy the variant
            * field mask above.
            */
            var nibble = (match == 'y') ?
                (randomNibble & 0x3 | 0x8) :
                randomNibble;

            /*
            * Ensure the nibble integer is encoded as base 16 (hexadecimal).
            */
            return nibble.toString(16);
        }
    );
};

ES6 sample

const guid=()=> {
  const s4=()=> Math.floor((1 + Math.random()) * 0x10000).toString(16).substring(1);     
  return `${s4() + s4()}-${s4()}-${s4()}-${s4()}-${s4() + s4() + s4()}`;
}

I adjusted my own UUID/GUID generator with some extras here.

I'm using the following Kybos random number generator to be a bit more cryptographically sound.

Below is my script with the Mash and Kybos methods from baagoe.com excluded.

//UUID/Guid Generator
// use: UUID.create() or UUID.createSequential()
// convenience:  UUID.empty, UUID.tryParse(string)
(function(w){
  // From http://baagoe.com/en/RandomMusings/javascript/
  // Johannes Baagøe <baagoe@baagoe.com>, 2010
  //function Mash() {...};

  // From http://baagoe.com/en/RandomMusings/javascript/
  //function Kybos() {...};

  var rnd = Kybos();

  //UUID/GUID Implementation from http://frugalcoder.us/post/2012/01/13/javascript-guid-uuid-generator.aspx
  var UUID = {
    "empty": "00000000-0000-0000-0000-000000000000"
    ,"parse": function(input) {
      var ret = input.toString().trim().toLowerCase().replace(/^[\s\r\n]+|[\{\}]|[\s\r\n]+$/g, "");
      if ((/[a-f0-9]{8}\-[a-f0-9]{4}\-[a-f0-9]{4}\-[a-f0-9]{4}\-[a-f0-9]{12}/).test(ret))
        return ret;
      else
        throw new Error("Unable to parse UUID");
    }
    ,"createSequential": function() {
      var ret = new Date().valueOf().toString(16).replace("-","")
      for (;ret.length < 12; ret = "0" + ret);
      ret = ret.substr(ret.length-12,12); //only least significant part
      for (;ret.length < 32;ret += Math.floor(rnd() * 0xffffffff).toString(16));
      return [ret.substr(0,8), ret.substr(8,4), "4" + ret.substr(12,3), "89AB"[Math.floor(Math.random()*4)] + ret.substr(16,3),  ret.substr(20,12)].join("-");
    }
    ,"create": function() {
      var ret = "";
      for (;ret.length < 32;ret += Math.floor(rnd() * 0xffffffff).toString(16));
      return [ret.substr(0,8), ret.substr(8,4), "4" + ret.substr(12,3), "89AB"[Math.floor(Math.random()*4)] + ret.substr(16,3),  ret.substr(20,12)].join("-");
    }
    ,"random": function() {
      return rnd();
    }
    ,"tryParse": function(input) {
      try {
        return UUID.parse(input);
      } catch(ex) {
        return UUID.empty;
      }
    }
  };
  UUID["new"] = UUID.create;

  w.UUID = w.Guid = UUID;
}(window || this));

The native URL.createObjectURL is generating an UUID. You can take advantage of this.

function uuid() {
  const url = URL.createObjectURL(new Blob())
  const [id] = url.toString().split('/').reverse()
  URL.revokeObjectURL(url)
  return id
}

I couldn't find any answer that uses a single 16-octet TypedArray and a DataView, so I think the following solution for generating a version 4 UUID per the RFC will stand on its own here:

const uuid4 = () => {
    const ho = (n, p) => n.toString(16).padStart(p, 0); /// Return the hexadecimal text representation of number `n`, padded with zeroes to be of length `p`
    const data = crypto.getRandomValues(new Uint8Array(16)); /// Fill the buffer with random data
    data[6] = (data[6] & 0xf) | 0x40; /// Patch the 6th byte to reflect a version 4 UUID
    data[8] = (data[8] & 0x3f) | 0x80; /// Patch the 8th byte to reflect a variant 1 UUID (version 4 UUIDs are)
    const view = new DataView(data.buffer); /// Create a view backed by a 16-byte buffer
    return `${ho(view.getUint32(0), 8)}-${ho(view.getUint16(4), 4)}-${ho(view.getUint16(6), 4)}-${ho(view.getUint16(8), 4)}-${ho(view.getUint32(10), 8)}${ho(view.getUint16(14), 4)}`; /// Compile the canonical textual form from the array data
};

I prefer it because:

• it only relies on functions available to the standard ECMAScript platform, where possible -- which is all but one procedure
• it only uses a single buffer, minimizing copying of data, which should in theory yield performance advantages

At the time of writing this, getRandomValues is not something implemented for the crypto object in Node.js. However, it has the equivalent randomBytes function which may be used instead.

The better way:

function(
  a, b               // Placeholders
){
  for(               // Loop :)
      b = a = '';    // b - result , a - numeric variable
      a++ < 36;      //
      b += a*51&52   // If "a" is not 9 or 14 or 19 or 24
                  ?  //  return a random number or 4
           (
               a^15              // If "a" is not 15,
                  ?              // generate a random number from 0 to 15
               8^Math.random() *
               (a^20 ? 16 : 4)   // unless "a" is 20, in which case a random number from 8 to 11,
                  :
               4                 //  otherwise 4
           ).toString(16)
                  :
         '-'                     //  In other cases, (if "a" is 9,14,19,24) insert "-"
      );
  return b
 }

Minimized:

function(a,b){for(b=a='';a++<36;b+=a*51&52?(a^15?8^Math.random()*(a^20?16:4):4).toString(16):'-');return b}

The following is simple code that uses crypto.getRandomValues(a) on supported browsers (Internet Explorer 11+, iOS 7+, Firefox 21+, Chrome, and Android Chrome).

It avoids using Math.random(), because that can cause collisions (for example 20 collisions for 4000 generated UUIDs in a real situation by Muxa).

function uuid() {
    function randomDigit() {
        if (crypto && crypto.getRandomValues) {
            var rands = new Uint8Array(1);
            crypto.getRandomValues(rands);
            return (rands[0] % 16).toString(16);
        } else {
            return ((Math.random() * 16) | 0).toString(16);
        }
    }

    var crypto = window.crypto || window.msCrypto;
    return 'xxxxxxxx-xxxx-4xxx-8xxx-xxxxxxxxxxxx'.replace(/x/g, randomDigit);
}

Notes:

• Optimised for code readability, not speed, so it is suitable for, say, a few hundred UUIDs per second. It generates about 10000 uuid() per second in Chromium on my laptop using http://jsbin.com/fuwigo/1 to measure performance.
• It only uses 8 for "y" because that simplifies code readability (y is allowed to be 8, 9, A, or B).

If you just need a random 128 bit string in no particular format, you can use:

function uuid() {
    return crypto.getRandomValues(new Uint32Array(4)).join('-');
}

Which will return something like 2350143528-4164020887-938913176-2513998651.

Just another more readable variant with just two mutations.

function uuid4()
{
  function hex (s, b)
  {
    return s +
      (b >>> 4   ).toString (16) +  // high nibble
      (b & 0b1111).toString (16);   // low nibble
  }

  let r = crypto.getRandomValues (new Uint8Array (16));

  r[6] = r[6] >>> 4 | 0b01000000; // Set type 4: 0100
  r[8] = r[8] >>> 3 | 0b10000000; // Set variant: 100

  return r.slice ( 0,  4).reduce (hex, '' ) +
         r.slice ( 4,  6).reduce (hex, '-') +
         r.slice ( 6,  8).reduce (hex, '-') +
         r.slice ( 8, 10).reduce (hex, '-') +
         r.slice (10, 16).reduce (hex, '-');
}