I wanted to tweak some stuff, so I installed the NodeJS platform on my server and had my own instance running in like an hour. Jitsi is the most mature webRTC lib out there, and believe me, I spent MONTHS researching.
No lol. But I spent about a year researching WebRTC and building demos, sadly which none came to purchase. I know all the big boys out there, and Jitsi is just simply the BEST.
Vonvo.com was using Tokbox since Meteor v0.6.5, and had a good run for 2 or 3 years. The site was eventually taken down, but we were featured in Devshop 8 and the Meteor Blog and had a fully working multi-person video chat infrastructure that saw quite a bit of production use.
I’ve been looking into creating a new telehealth product recently, and will probably be going with the opentok-react library.
The biggest challenge for WebRTC is device independency. If you’re just developing for web browsers, it’s fairly easy, since there are a lot of libraries out there that do the heavy lifting. But if you also target mobile devices, it becomes a challenge, especially with iOS.
We started with OpenTok / Tokbox, but their Cordova support was non-existent and the very old Cordova library that somebody at their team developed years ago was very buggy. We patched it and made it work, but all of that hazzle wasn’t worth the quite high price tag Tokbox is charging if you are really using this in production. It may work for a B2B use case, but not for C2C for which we use WebRTC.
Most of the other service providers have vanished or re-focused on building products instead of services. So, in effect, we cut our connection to Tokbox and developed our own service based on OS libraries we patched quite a bit to make everything work.
Jitsy is looking cool, though, we’ll have a look at it.
Jitsi looks excellent, but it appears to require you to install the back-end on your own server. Ideally I would like to have something where the dev ops are handled by a service provider.
While not ruling out Jitsi, I think I’ll have a look at Temasys next.
Go copy the source of that page in to a new .html doc on your desktop. Bam. Meetings. You can use Jitsi Embed on any page you want, and change rooms with a parameter…
I’ve been trying out Temasys with encouraging results. Dropping the few lines of code from their codepen demo into the client, quickly results in a working webRTC videocall capability.
I thought Temasys’ backend platform (for multiuser meetings with advanced features and whatnot) is a paid service, except for a very limited free tier whereas jitsi has no limitations.
Yes, Temasys is a paid service with a limited free tier to support development. At the same time, per a Jitsi-Meet Github issue thread, Jitsi requires the developer to install Jitsi back end code on the application server, which I am seeking to avoid for purposes of my app.
I asked specifically on that thread if I could just drop the html/javascript code provided above into the client side and get videocalls working, but the response appears to indicate that this would not work until Jitsi back-end code is installed on my app server.
If convenient, please have a look at that thread and let me know if they missed anything or if there is an approach to using Jitsi that does not require a back-end to be installed on my app server.
WebRTC always needs a back-end, for the initial handshake to setup the connection (ICE protocol). When everything is set-up, communication happens directly on a P2P basis.
However, you would also need to implement a media server if you want to support conferences with many participants. It is possible to do it without, but that causes a lot of network traffic and load on the devices.
Yes. Am I correct in thinking that a service such as Temasys can provide that back end? In development I added Temasys code to the front-end only and was able to set up a working audio-video call between two computers.