When I run jitsi-meet in dev mode, what is running locally?

I downloaded the source and am running jitsi-meet on, and I’m assuming that is only the meet application and a basic web server (when looking at the architecture diagram: Architecture | Jitsi Meet).

I’m assuming it is using a remote prosody server, but I’m wondering if that’s hard-coded and/or what ports it is running over. I assumed port 5280 (based on Self-Hosting Guide - Manual installation | Jitsi Meet), but looking at tcpdump for that port didn’t show any outbound traffic.

If the development instance is running xmpp/jicofo/vb locally, that’s fine, I just didn’t think that’s how it would work.

jitsi-meet, prosody and jicofo are on the same server in a typical setup. Prosody’s 5280 is set as an upstream in nginx conf

The make dev just runs a proxy server locally so you can use the local UI code and everything else is used from a remote server. It works by default with a server we have setup with limited capabilities, but you can always setup and use yours:

1 Like
const devServerProxyTarget
    = process.env.WEBPACK_DEV_SERVER_PROXY_TARGET || 'https://alpha.jitsi.net';

Exactly what I was looking for – thanks!

It looks like this alpha.jitsi.net has a 5 minute time limit on conferences, for example… not a problem for me or this level of testing.

and a follow-up question:

If my project actually gets to a point where I want to go live with a new frontend app, but I don’t want to deal with hosting any of the backend components, I’m assuming I can just point the app to an 8x8 Proxy Server (paid solution)? Or am I oversimplifying that?

Just to confirm that jitsi documentation was not lacking (I just wasn’t reading the right things), this is where the devServerProxyTarget is described in the Developer’s Guide:

(my bad)

You need to either host your own deployment, instructions on how to install it: https://jitsi.org/qi. Another option is to use JaaS: https://jaas.8x8.vc/.