Accessing Arbitrary Ports

If you already have a server running on localhost listening on a port, you can access it through the notebook at <notebook-base>/proxy/<port>. The URL will be rewritten to remove the above prefix.

You can disable URL rewriting by using <notebook-base>/proxy/absolute/<port> so your server will receive the full URL in the request.

This works for all ports listening on the local machine.

With JupyterHub

Let’s say you are using a JupyterHub set up on a remote machine, and you have a process running on that machine listening on port 8080. If your hub URL is myhub.org, each user can access the service running on port 8080 with the URL myhub.org/hub/user-redirect/proxy/8080. The user-redirect will make sure that:

  1. It provides a redirect to the correct URL for the particular user who is logged in
  2. If a user is not logged in, it’ll present them with a login screen. They’ll be redirected there after completing authentication.

You can also set c.Spawner.default_url to /proxy/8080 to have users be shown to your application directly after logging in - without ever seeing the notebook interface.

Without JupyterHub

A very similar set up works when you don’t use JupyterHub. You can construct the URL with <notebook-url>/proxy/<port>.

If your notebook url is http://localhost:8888 and you have a process running listening on port 8080, you can access it with the URL http://localhost:8888/proxy/8080.

This is mostly useful for testing, since you can normally just access services on your local machine directly.

From Notebook Extension

If you have a client side extension for the classic Jupyter Notebook interface (nbextension), you can construct the URL for accessing your service in this way:

define(['base/js/utils'], function(utils) {
  // Get base URL of current notebook server
  var base_url = utils.get_body_data('baseUrl');

  // Construct URL of our proxied service
  var service_url = base_url + 'proxy/' + port;

  // Do stuff with your service_url
});

You can then make HTTP / Websocket requests as you wish from your code.

From JupyterLab Extension

Accessing your service from a JupyterLab extension is similar to accessing it from a classic notebook extension.

import { PageConfig } from '@jupyterlab/coreutils';

// Get base URL of current notebook server
let baseUrl = PageConfig.getBaseUrl()

// Construct URL of our proxied service
let serviceUrl = base_url + 'proxy/' + port;

// Do stuff with your serviceUrl