Good news, everyone: after spending a little over 6 months in beta, Mux Real-Time Video is now in General Availability!
That means that you can develop all your amazing real-time video applications with confidence, knowing you're building on top of a stable foundation from the team at Mux.
Starting today, Real-Time Video is accessible to everyone, directly from the Mux Dashboard, so you can just sign up and get building1. All the documentation you need is available in our docs site, including details on how to get started building real-time video experiences into your web, React, iOS, and Android applications.
When we started building Mux Real-Time Video, we were at the peak of the pandemic, and new video use cases and video-centric startups were popping up all over (I've never seen "Metaverse" in so many pitch decks!). As we've moved through that period of uncertainty and put Real-Time into more peoples' hands, that creativity has exploded, leading to all sorts of new applications and experiences.
We've seen hybrid fitness experiences, innovative video-conferencing applications, content creation flows, collaboration experiences built into a cacophony of applications, and new events platforms, all enabled by Real-Time Video.
But Video is changing. Video streaming in 2022 isn't just about a single real-time, live, or on-demand video experience. It’s about a more complete video strategy that transitions between all these types of video — and that strategy is at the core of how we've built Real-Time Video.
Real-Time Video is a central component of Mux Video, which means it takes a single API call to broadcast any real-time session and have millions of viewers watching with 4 to 7 seconds of latency on a low-latency live stream. And of course, just like with any other Mux live stream, the on-demand assets are immediately available for viewing. If you ask me, I think those streams are best consumed in the shiny, newly GA'd Mux Player… but I may be biased.
Well, it takes a lot of time to delete all those BETA icons we put on the website… Just kidding!
You might notice that Real-Time Video looks a lot like it did when we announced it earlier this year, and that's deliberate. The team has been mostly focused on things that you can't see. In software, this is generally referred to as "non-functional requirements". This mostly includes making the product scalable, reliable, and secure.
For the last 6 months, the team has been hard at work behind the scenes ensuring that Real-Time will scale as your business grows and that it will work reliably across a large footprint of devices, even in challenging network environments. If you're interested in a more comprehensive list of our non-functional changes, check out this tweet thread from our amazing development team. Here are some of the highlights:
- Transparent multi-region support
- Scaled the service over 10x to support usage through the beta
- Refined our scaling behavior to seamlessly scale as usage grows
- Integrated QoS metrics collection
- Comprehensive load testing of the biggest spaces
- Lots of 🐛 squashing, and much more
We've made it even easier to integrate Real-Time Video into your application. While we've made lots of smaller changes to our SDKs to make them more modular and easier to develop against, we've also made one larger change that we're really proud of: automatic subscription management.
When building a real-time video application with lots of participants, it's often hard to know which participants' video and audio you should be subscribed to. Historically, this has meant complex management logic for a developer, and it's easy to run up against limits. Our automatic subscription mode is designed to make this easy. We use speech detection and other contextual hints to understand who the most important participants are to be subscribed to and to automatically manage those subscriptions as needed when conditions change.
If you're building something more unique and don't want automatic subscription management, it's easily turned off. Check out the guide for more details.
When your participant leaves a session, they're now able to share their feedback about the quality of the audio and video they experienced. This data is really important not only for us here at Mux but also for you, the builder of your application.
This feedback is scored on an Absolute Category Rating scale, a common scale for rating subjective audio and video quality.
Initially, these scores are for Mux internal use, as we're still finalizing the API for getting feedback from users back out of the platform. Be on the lookout for the upcoming Real-Time Sessions API.
When we first built Real-Time, we built an open-source meeting application, Mux Meet, so that we could dogfood our platform. We use Mux Meet every day on teams across Mux, for large and small calls.
We've continued to iterate on Mux Meet, to make it a true reference implementation of all the best features of Mux Real-Time Video — and yes, that includes ACR scores, automatic subscription management, and broadcasting features. We've even recorded a getting started video, which will have you up and running in no time.
We've got lots of new features and functionality in the pipeline for Real-Time Video. Keep your eyes on the blog and subscribe to the Mux changelog to make sure you don't miss out on our upcoming innovations.
But we're also a long way from being done, and our roadmap isn't set in stone. If you're building something really cool that has some special requirement we aren't currently meeting (get it?), we'd love to hear from you!Get started with Real-Time Video
We can't wait to see what you're going to build with Real-Time Video. If you need some inspiration, you should also be sure to check out the original announcement blog post.