View the Project on GitHub Stuk/require1k


Copy and paste or download:

<script>R=function(e,n,v1){function t(e,o,u,a){if(e.g)return o(e.e,e);var c=e.g=e.l,f=new XMLHttpRequest;f.onload=function(i,l){function s(){l--||o(n,e)}200==f.status||e.t?(i=[],(e.t=e.t||f.response).replace(/(?:^|[^\w\$_.])require\s*\(\s*["']([^"']*)["']\s*\)/g,function(e,n){i.push(n)}),l=i.length,{t(r(e.l,o),s,"."!=o[0]?c+"/../":n,o)}),s()):u?t(e.n=r(u+="../",a),o,u,a):(e.e=f,o(f,e))},e.t?f.onload():("GET",c,!0),f.send())}function r(e,n,t){if(e.e)throw e.e;return n?(f.href=e,i.href="."!=n[0]?"./node_modules/"+n:n,t=i.href+".js",f.href="",u[t]=u[t]||{l:t}):e.n?r(e.n):(e[c]||(e.f||a("(function(require,"+c+",module){"+e.t+"\n})//# sourceURL="+e.l))(function(n){return r(r(e.l,n))},e[c]={},e),e[c])}function o(e,n){t({l:"",t:""+e,f:e}:r("",e),function(t,o){try{e=r(o)}catch(u){t=u}n&&n(t,e)})}var u={},a=eval,c="createElement",f=e[c]("base"),i=e[c]("a");return e.head.appendChild(f),c=e.querySelector("script[data-main]"),c&&o(c.dataset.main),c="exports",o}(document);</script>

View the uncompressed version on Github.

Vote/comment on HN

CommonJS require for the browser in 1k, with no build needed

This project implements a minimal, and yet practically useful, CommonJS/Node.js require module loader for the browser in under 1000 bytes.




In your modules you can use the require function, exports object and module.exports object as you would in Node. There are two ways to kick off the require system:

script data-main attribute

<script src="require1k.min.js" data-main="./index"></script>

On start require1k will search for the first <script> tag with a data-main attribute. If found the named module will be resolved against the location of the html file, asynchronously loaded and then executed.

global R(func, callback)

R(function (require, module, exports) {
    var index = require("./index");
    exports.hello = "World!"
}, function (err, exports) {
    if (err) {


Require1k also adds a global function, R, that accepts either a function or a module ID and an optional callback. When given a function it will load all the dependencies of the function and then execute it. If there was an error the callback gets passed the XMLHttpRequest object that failed as the first argument. It also gets passed the exports of the function, which you may find useful.

global R(moduleId, callback)

R("./index", function (err, index) {
    if (err) {
    // use the exports object as needed

Alternatively you can provide a module ID and a optional callback. The named module is resolved against the location of the HTML file. If there was an error the callback gets passed the XMLHttpRequest object that failed as the first argument, otherwise the exports of the module are passed as the second argument. The callback may be called synchronously if the module is already in the internal module cache.


The point

The short answer: because it's a fun hack.

The long answer?

Using <script> tags to manage the dependencies of your "complex" web app makes for an unmaintainable mess.

Yet that's what most people are doing today, and even what a large number of articles currently being published still recommend. People are writing large and sophisticated apps and then wasting their time managing dependencies by hand! There's just no need for it.

Package management ensures you have a system for adding and updating libraries to your project. It means that you can easily check if your libraries are out of date, and see exactly what libraries your project depends on. It means that you don't have to go spelunking around the web to find the site where you can report an issue or download an update for this random file you have in your js/ directory. But module management is even more important.

Module management is where you actually say "load this code for my app". And when you do this with a (multitude of) <script> tags on your page you can no longer isolate your code, and say "this is what this bit of code depends on". Every library you load is global. And what happens when you write some JavaScript that needs an extra library? You have to go back to this completely different file, and add something to the HTML. This is nonsense.

What each part of your app depends on should be explicit. It should not require editing an unrelated file just to depend on a third party module.

If you're writing a web app of any significant size today, you owe it to yourself to manage your dependencies correctly.

This is what I recommend:

That's it.

(And yep, I've noticed the irony of saying this while telling people to embed require1k on a page. But this is just a joke. Kind of.)


This project was largely inspired by working on Mr, a CommonJS module loader and companion bundler that definitely isn't 1k.

Thanks to Tom Robinson for writing the original version, and Kris Kowal for teaching me so much.