Jitsi with high number of participants


I am using Jitsi software and I want to be able to host a conference with more than 100 users. Currently, I already set up multiple JVB’s, enabled Octo, and successfully distribute all traffic to multiple servers for a single conference. For our use case, we need to support more than 100 users with video. This conference will have one main speaker but randomly, it’s possible for another user to open video/audio and answer to the main speaker. I tried configurations like channelLastN but that’s not exactly what we want since it disables video based on activity. We are trying to implement pagination for the conference and only get the video/audio streams of the users for the active pagination page. Is there any easy way to implement/support this on JVB part? If we can only receive the video of users on the active pagination page, it will be easier for us to support more users I think.


1 Like

Not sure actually, @bbaldino is this supported at the moment? Can you request video for A, B and C and say no video for D, E, F or the rest …

No, not right now. We have some work we want to do here to make this possible, though. No timeline yet.

Thank you for your response. We kinda need this as soon as possible. Do you know someone that can help us develop these kinds of features or maybe help us to optimize our setup as a consultant?

Thanks for your work on this. This would give you an amazing amount of control to give the possible experience to the viewers. You all do an amazing job!!

@bbaldino Thank you for the information. We have a custom UI with lib-jitsi-meet where we are using setLastN in combination with selectParticipants in order to select specific feeds (according to who is visible on the screen, in order to do pagination via scrolling, more info in Being able to select remote participants for high res and participants for setLastN separately).

We noticed that some times, even if we the mentioned lib-jitsi-meet API calls, some participants cannot see other participants that should be visible. We believe it is related to Last N being changed with speaker audio activity, or because we are changing last N and select participants multiple times during the conference. Is that related to your comment? Will it be possible to do this in the future?

Thank you very much in advance we really appreciate and admire the work that you do in Jitsi development.


I thing pagination feature is very crucial for benchmarking jitsi with other conference applications especially with respect to client cpu performance and data consumption.

At that moment clients decodes the video streams that is not shown.

is there any update about this feature :slight_smile:

Yes I believe the work for this either just got merged or will be merged soon.


Can you confirm us when this feature is merged?

1 Like

Hi @bbaldino, any news about this?

Pretty sure this is in now, you can read about the new message format here.


Sorry for the ignorance here but if we wanted to implement this new method, how would we go about using it. Can we send this information across the iframe API or would we need a custom module to utilize the new method?

@jallamsetty can probably best answer that.


@jallamsetty talk to me. I’m in display heaven! :smiley:

oh this is far more exciting than a toolbar or a background effect IMO.

Me sitting here, holding my breath, watching @jallamsetty type a life-changing message… :joy:

The work needed in lib-jitsi-meet for translating the signaling messages for setLastN(), selectParticipants() and setReceiverVideoConstraint() to the new format is done and is behind the config.js flag useNewBandwidthAllocationStrategy. However, work is still under progress on the jitsi-meet side of things for supporting pagination. After that is done, we will expose it through the iframe API.


I was over here holding my breath and screaming at the monitor: “Girl, use speech-to-text if your fingers won’t let you type fast enough, before I turn blue!!!” :joy:

Thank you! Thank you!! Thank you!!! For this awesome news - and work, dev team!

1 Like

@johnny256 You should be able to use the setReceiverConstraints method directly instead of using setLastN and selectParticipants. If the endpoints are selected, those endpoints should be prioritized first overriding the natural speech activity order of the endpoints.
I have added the documentation for this new method in the handbook.


This is awesome, thank you so much for all the work in this new feature and also for the new documentation. This level of control on each client is amazing. It will help a lot with our use cases.

I appreciate very much the support not only jitsi-meet but also lib-jitsi-meet for custom implementations. The dev team is very flexible to the needs of the community.

Once again, thank you very much @jallamsetty and all the Jitsi team!!