Setup for an online escape game


first of all I want to thank all developers for creating this fantastic software and making it open source. It enabled me to stay in touch with friends and family during these hard times.

I’m in Germany and we’re on our second lockdown now that will last at least another four weeks. During the first lockdown my wife and I performed a live quiz show from our living for a bunch of friends using OBS and Jitsi. Now we want to make something more sophisticated - an live online escape game.

We thought of something like this: We invite everyone to a special conference room to do another quiz. After everyone joined an actor (a.k.a. me) comes into the chat. I’m just arriving at home from shopping groceries and my wife has been kidnapped. The kidnapper left some clues and the participants have to help me finding her within one hour.

I’d have two cameras strapped to my body, one pointing at my face and one bodycam pointing forward. I’ll also wear a headset to talk to the participants. A producer (a.k.a. my wife) sits in the basement and controls the main video feed in the conference. She can switch between the live feed of my bodycam, some prerecorded videos (with audio), static images and maybe a map with my current position. These are the clues and riddles that have to be solved.

My video and audio feeds should be connected via an LTE link so I can freely move around the house, garden or maybe even drive two minutes by car to a small forrest nearby. It would be nice to use two smartphones but I could also put a battery powered computer with an LTE uplink into a backpack.

A nice bonus would an audio connection to the producer that only I can hear.

I made a little POC with compositing the video feeds in OBS and sending it to a Jitsi server but that adds some large latency to everything as the video feeds go to the OBS on my home computer and then upstream to the server. Also it doesn’t handle audio very well. Instead I would like to stream the ‘actor cams’ and the content from the producers computer directly to a Jitsi server (I can set up an own server for that) and having the producer remotely control the video and audio composition on the server.

This is how the interface should look like to the participants:

Has anyone done a setup like this or point me to some software or setup instructions for this use case?


Found a solution:

I‘m using the external iframe api to create a conference page for the participants. This includes some javascipt that subscribes to the displayNameChange event.

Depending on the nickname of the producers stream (I’m just adding a symbol like ^ to the end) it uses the setLargeVideoParticipant command to change the source of the large video.

The producer has a separate page that includes some buttons (‚show bodycam‘, ‚show producer obs virtualcam‘, ‚show facecam‘). These call the displayName function to change the nickname which in turn leads to the participants windows toggle the large video source.

The face and body cam are just two smartphones with the jitsi app streaming to the same conference.

Apart from having the obs source, the face cam and the body cam as three participants in the film strip it does 90% of what I need. If I could hide the body cam and the obs stream from the film strip it would be perfect but this way it‘s good enough for a one-time-event.

BTW: I tried using chat messages first to remote control the participants conference window but it seems you don‘t get any messageReceived events when the chat button is hidden.