LiveKit Community

Looks like I can manually create the tracks and pass them in as FastConnect options? Seems strange to not be able to just setup the participant before connect with the same patterns used in the meeting.

Yes, you can use FastConnect options, which use pre-created LocalAudioTrack or LocalVideoTrack as room to connection parameters.

that's a good point Jesse, we'll take it into consideration as an improvement in the future

<@U0400GH4DBQ> is there something in particular you’re interested in learning about? We’ve talked about doing video tutorials, but it helps to understand what we should focus on. Is it getting started quickly (with OSS or Cloud)? Is it leveraging a particular feature/capability? Is it an end-to-end example of building something specific (i.e. a meeting app, twitch-inspired livestreaming app, etc.)?

Yes, I would love if there is a video tutorial on Livekit to create a one-on-one Video meeting and a Livestream as seen in TikTok.

I use Flutter.

I would really appreciate this if it is possible and besides there is no one single video tutorial on the internet about Livekit.

I think this will really help a lot of person like me

Thanks

At the risk of embarrassing myself, could you link me to an example on TikTok? :see_no_evil:  

1) I want the user of the app to be able to Livestream via the App, 

Example:
<https://vm.tiktok.com/ZMYCA4UGB/|https://vm.tiktok.com/ZMYCA4UGB/>
<https://tango.me/juicy55|https://tango.me/juicy55>



2) I want user to be able to engage in one-on-one meeting session or with a group of user.

E.g Like WhatsApp Video Call or Voice call


Am using Flutter and am definitely using the Livekit Cloud

<@U01KZ0SCDKL> please how do I proceed with step by step instructions

Hello <@U0400GH4DBQ>, maybe you've already seen this, but in case not, I know this isn't a Video tutorial on how to use it but it is a working example on how to integrate a Flutter application with LiveKit.

Flutter example: <https://github.com/livekit/client-sdk-flutter/tree/main/example>

<@U032TUAL4KT>, hey Pat, i just clone the repo, but am quite confuse, do i need to alway generate token at the LiveKit Cloud for usage?

Yes every users needs their own token to join a room

<@U0400GH4DBQ> you can use one of our server SDKs to do this on the backend for every user joining a room.

<https://docs.livekit.io/concepts/authentication/|https://docs.livekit.io/concepts/authentication/>

okay, I'm getting a hang of this, but I need clarity in this aspect:

am using Fluter as my Client
Nodejs as my server
and LiveKit Cloud for the over operation.

my question its, how do I connect my Nodejs server with LiveKit Cloud?

can I host my Nodejs server in LiveKit Cloud?

It depends on what you want to do with your node server. 

You don’t explicitly need to connect them. At a minimum, your node server simply needs to use the JS server SDK to generate a token, which your flutter client app uses to connect to LKC.

We do not currently, nor do we have plans to, host third-party code in LKC. There are other services like Vercel, Railway, Firebase, Supabase, which can help there.

It uses about 30MB of memory at idle. What memory footprint are you trying to get it down to?

There aren't really flags are available that can get it to use less

is there an easy way to do this <@U02GAC58A79>?

I’ll defer to Duan for flutter/dart. For nodejs there is no way to publish any media currently but if you want to publish media serverside, you can check out our Golang server client: <https://github.com/livekit/server-sdk-go|https://github.com/livekit/server-sdk-go>

It’s also possible via client-side JavaScript using WebAudio, check out this file for how to do that: <https://github.com/livekit-examples/spatial-audio/blob/main/src/controller/JukeBoxProvider.tsx|https://github.com/livekit-examples/spatial-audio/blob/main/src/controller/JukeBoxProvider.tsx>

Thank you a lot, and I am waiting patiently for any feedback about dart

hey <@U050MGLJU9X>, this is a feature about insertable streams for flutter-webrtc (use custom files or player source to create audio track), this is indeed in our roadmap, but recently we are still implementing End to End Encryption and preparing new ones flutter meeting app. when I finish these things (it will take about two months), I will try to push the full insertable streams feature for flutter-webrtc.

&gt; is there an easy way to do this <@U02GAC58A79>?
<@U01KZ0SCDKL> It’s not easy, but I think it will be easier than e2ee. :thinking_face:

<@U02GAC58A79> thank you for the response, and good luck with your work.

Please help me the document setup livekit server with Go code.

you need to use a custom template for this: <https://docs.livekit.io/server/egress/custom-template/>

I'm not familiar with GPUImage but as long as you implement `VideoRenderer` protocol you can receive frames and render however you want .

Thanks, I will try some solutions and your solution.

I've heard some folks sharing that they did it. It should be possible to transform the audio prior to publishing it.

I diged deeper and it seems that WebRTC has noise reduction implemented in the protocol.

I hope thats good enough. We want to have it well for outdoor and sports situations. Its a bummer that krisp is only mac and windows.

Yeah, the standard reduction isn't bad. but it's def not as good as Krisp. we are exploring an integration with them. I believe they have native SDKs too

Even for iOS/Android? We would be happy to be a paying early adopter/alpha tester.

Hello dear forum membres
We have one trouble when develop on IOS -
when phone locked(fingerprint or pin-code) and we receive VoIP push - we connect to room and receive and transmit video
but we can’t manage by voice - and can’t hear anything
It is like this case:
<https://github.com/flutter-webrtc/flutter-webrtc/issues/1005#issuecomment-1188742866>
May be some one knows - how switch on voice after unlock iPhone?

Some requests are working fine and returning `200` status code

<@U04GJGHGACS> i think we’ll need some more details here to be able to help. Do you have any code you can share? Are you using LiveKit Cloud or self-hosted? Any particulars around which requests are failing and which are succeeding?

Well it seems like the 500 failed requests were because of some rate-limiting on the server side, if I make the requests go synchronously one after the other then I don't get failed request

but if i let them go Asynchronously in parallel then I get these failed requests

is there a way to configure/disable the rate limits?

which environment are you using? if Cloud, we do have a 1k request/min limit on APIs

<https://docs.livekit.io/cloud/quotas-and-limits/#api-rate-requests>

you’re making requests from a local client to a local livekit server?

you'd want to look at the server logs to see what that failure is..

The server logs don't produce something usefull

<https://github.com/moham96/livekit-test>

it seems that when setting the metadata of a participant to something other than an empty string I get the failed requests

I'm not sure if it is connected to rate limiting or something else

there should not be any rate limits there..

how large is the metadata you are trying to update? and what does your server deployment look like?

image.png

you should be able to pass your own sorting function in as a prop to the Stage!

also have a look at the new react components, as we will deprecate the livekit-react packages soon,
the new ones live here: <https://github.com/livekit/components-js>

on mac, you can just run `brew install livekit-cli`

However, I was able to do that following some tutorials.

I can stream audio but the egress functionality that handles recording fails

but it doesn't upload the recording S3 Bucket

you can find it using `livekit-cli list-egress`

I am just reporting based on the feedback received from the developer who tried to set it up

This is why I am in search of someone who is experienced to handle it.

Also willing to pay for it

It also only seems to report one of four values if changed to a float. { 0, 0.125, 0.25, 0.375 }

Hey Greg, you’re right, I’m going to release a fix later today :slightly_smiling_face:

it's not implemented yet, but we are working on it!

ok ,thanks ....this is one of the requirement . Could you please update us once this is achieved ?

you can follow this issue for updates: <https://github.com/livekit/livekit/issues/1565>

hi <@U01KZ61GV5Z> , i can see this issue got closed ? could you please help me how can i create token for a user who can publish ONLY  AUDIO and subscribe audio video both ?

Hello, I would like to ask if there are any plans to improve importing CSV files from the dashboard.

In the session page, I can download a CSV file from the participants table, but all values in the "published tracks" column are written as "[object Object]."
I wonder if there will be a way to obtain the stringified values of that array, such as "microphone, camera, screen share," in the future.

Thank you in advance.

Yup, we’ve been talking about it! cc <@U03H6MNQ9G9> 

hi joshua!
that shouldn’t be a problem at all, you can just use the `livekit-client` package directly in any vanilla js app.

in case you’re looking for ready made designs, the new components packages (<https://github.com/livekit/components-js>) also include a css only package `@livekit/components-styles` which you could use around some custom components if you don’t want to use any ui framework.

Cool, thanks. I was also thinking of just co-opting the `meet` app, which seems like it can work if I just embed it within another app

Any reason not to do that? Seems like token auth will do the job I need, which is preventing people from joining any room

embedding the meet app should work just fine!
It’s built on nextjs, so the auth provider runs on the server. you’d probably have to separate that from the rest and host your token end point some where else.
 Depending on your needs it might make sense to adjust the token auth path to only allow your users as currently any one can join. But that of course depends on your use case!

Thanks.I thought maybe to require having rooms be created before joining. Would that be enough to prevent anyone from using it?

sure, if you disable auto room create and set the `roomCreate` permission on the user’s token to false, that should work, too!

you’ll then have to use the server-sdk to create the room specifically (note that they will timeout/close after a while with no participants in them) <https://docs.livekit.io/server/room-management/#createroom>

Hey David, it’s definitely on the roadmap, but most likely something we tackle in H2 2023. Perhaps <@U02GAC58A79> can give some guidance on how you might integrate it yourself in the meantime?

<@U046B0LNLD7> If you want to try to do it yourself, please take a look at this repo <https://github.com/cloudwebrtc/go-sip-ua/tree/feat/webrtc2sip-gateway/>. I created an experimental branch of webrtc2sip last weekend, which is used for two-way communication between webrtc and sip client. Currently, it can communicate with port-sip and linphone, but docking with livekit server still needed some modifications, such as changing one side of b2bua to livekit server sdk, when multiple audio tracks come from the livekit room, you need a mixer to mix the audio tracks into one and send it to the sip side.

Please check this directory <https://github.com/cloudwebrtc/go-sip-ua/tree/feat/webrtc2sip-gateway/examples/b2bua/b2bua>

`rtc.go` and `udp.go` correspond to `webrtc` and `sip` media track handlers.

These lines of code are part of the media stream bridging of the two protocols
<https://github.com/cloudwebrtc/go-sip-ua/blob/feat/webrtc2sip-gateway/examples/b2bua/b2bua/call.go#L175-L202>

thanks for the input <@U01KZ0SCDKL> and <@U02GAC58A79> this is really helpful!

I’ll see if we can move the SIP integration back on our roadmap, I think it makes most sense to wait until you’ve done the integration

Yes and no Depends on your requirements.

As far as I understand the cloud version supports rooms of any size but in the open source version this feature is not there. 

So you are basically restricted by the number of consumers that one of your servers can handle.

We have been using livekit for sometime now as conferencing platform but we also need 1-to-many now. We are in the process of using the server-sdk-go to create a kind of a bridge to interconnect rooms. This will allow us to scale according to our needs and use as many servers as needed with the open source version as well. But if course this needs some custom development.

<@U051Y36UJ3E>, here is a link to a livestreaming example if you are interested in creating your own livestreaming setup: <https://github.com/livekit-examples/livestream>
If you have more questions we are happy to help. :slightly_smiling_face:

<@U051Y36UJ3E> the OSS server does support one-to-many broadcasts. You can check out benchmarks here:

<https://docs.livekit.io/oss/benchmark/#benchmarks|https://docs.livekit.io/oss/benchmark/#benchmarks>

Does LiveKit Cloud support one to many?  Or will that come soon?

<@U03S81VRSHF> Yeah it does! <https://livekit.io/cloud|Cloud> supports up to 100k simultaneous viewers on a stream.

<@U01KZ0SCDKL> Thank you for your answer. I just look up another framework to replace Kurento. If Livekit can do, i'll try :D

<@U051Y36UJ3E> we’ve met other folks that have switched to LiveKit from Kurento, so I think it should work well. :blush:  

Screenshot 2023-04-06 at 16.04.39.png

hey Max, currently there isn't a way to retrieve this data. we do have plans on making the analytics data accessible via APIs though

<@U01KZ61GV5Z> ah ok yeah that would be handy.
In that case I was wondering if the pricing calculator (<https://livekit.io/pricing/calculator>) is completely reliable, and is everything that affects cost listed on there?

The pricing calculator is just a high level estimate. If you have a specific workload in mind, feel free to DM me and I can help you figure it out

LiveKit can help with bringing any external video stream into the conversation. As long as you can render the content on the server side, it should be straight forward to publish it as a video stream (tagged with `TrackSource.SCREEN_SHARE`)

Do you already have a way of broadcasting a pdf/doc on the server side?

Hi David, thank you for the reply!

No, we have no way of broadcasting a pdf/doc on the server side. The goal is to do it on the frontend

Our user case is as follows:
Two users in a video call, the documents are only accessible by one of them in the platform and when clicking on a certain document, it is broadcasted in real-time to both users in the video call, being possible to scroll through the document and even highlight and/or write notes.

I don't think LiveKit can help with the document sharing part.. it'd be best if you stored that in cloud storage and opened them locally with browser components

What if the canvas of e.g. pdf.js is used to generate a stream? And capture that stream using the HTMLCanvasElement: captureStream() method?
Then this stream could be published to the livekit with the publishTrack function?
This is what we are trying to do right now.

ah yeah, for PDFs that would work well. you can definitely publish any client-side Canvas to livekit

that's a clever way to avoid using screen sharing!

Hi Quedo, psrpc is <https://github.com/livekit/psrpc/|here>. we are using it extensively within livekit for server to server communications

Hi, are there any docs for the code or logic so that I can understand how it works internally?

not currently. we do plan to add docs detailing the protocol/internals. but the high level is you can think of pubsub messages sent back &amp; forth, with things like retries/etc builtin

Ok thanks. I'll try to tinker around on my own.

Yes, our JavaScript SDK works and is tested with all major browsers including Safari

i’m not really familiar with the windows cli but my guess is you need to remove those `\` characters

yeah i was wraong , i fixed it but im using livekit custom room to connect and test but it dosnt connect

<@U052P6F5449> please thread your messages :pray:

<@U052P6F5449> writes:
&gt; I need help, when using JS SDK it doesn’t import in browser.
&gt; Any CDN link for JS-SDK library ?

are there any CDN or any other method to do the task?

Hey <@U052G0R3WL8>, we’ll be dropping a sample app soon which demonstrates one way to do real-time transcription. The technique we’re using leverages the Go server SDK piped to Google STT.

The Go and Rust SDKs are currently the only ones that have access to WebRTC implementations on the backend. Eventually we want to bring that to the other platforms, but for now the aforementioned are your best bet. 

Another technique is to hook into the audio track on the client, send it to a transcription service, then send the transcribed text to whomever needs to receive it via data messages. This way would allow you to do this in JS/TS.

Symbl wrote a blog post about how to do this:

<https://symbl.ai/blog/how-to-build-a-video-conferencing-application-with-livekit-and-symbl-ai/|https://symbl.ai/blog/how-to-build-a-video-conferencing-application-with-livekit-and-symbl-ai/>

Hello <@U01KZ0SCDKL>, thank you so much for your reply. I wanted to ask how you would collect audio directly from the client side (which is subscribed to the livekit server's audio) itself, if that makes any sense.

Maybe something to add to code like this.
```room.on(RoomEvent.TrackSubscribed, (track, publication, participant) =&gt; {
    console.log("Participant Subscribed");
    attachTrack(track, participant);
});                                   function attachTrack(track, participant) {
  const v = document.getElementById("them");
  track.attach(v);
}                                                               ```
Once again, I greatly appreciate your help.

perhaps <@U02EU2ZPTQU> or <@U01UEMQ8C3F> might have done this in the past?

hi <@U052G0R3WL8>,
it depends on what you want to do with it. we have a helper function for audio/volume sampling as part of the client-sdk-js that’s called `createAudioAnalyzer`, you pass in a livekit track and it returns an audio analyzer node attached to that track among other things.
If what you’re looking for is more application specific, you can look at the code of that helper to figure out who do get it to work for your specific use case
<https://github.com/livekit/client-sdk-js/blob/main/src/room/utils.ts#L311-L365>

Hey Piyush, there are a couple of ways of achieving this:

1. use separate rooms for stage &amp; backstage.
2. use <https://docs.livekit.io/client/publish/#subscription-permissions|subscription permissions> on each publisher, explicitly indicating who is allowed to subscribe to the publisher's tracks

Team, any rough ETA for the Unity native SDK release?