How Multi is building next-generation live collaboration with Zoom Video SDK

Multi set out to build a collaboration tool that aligned with their own philosophy, one that’s shared by other high-performance software teams: Spend time building together, invest in great asynchronous follow-through, and have as few other meetings as possible. 

Sessions in Multi are built around multiplayer app sharing. When someone shares an app in a session, everyone can make changes with shared cursors, drawing, and even shared control. Colleagues can work in each other’s apps like Xcode and Terminal as if they were open on each person’s own computer. 

After a session is over, AI-generated summaries capture insights and action items, which helps push that knowledge into the places everyone on a team needs to see them, like docs, messages, and even the apps they’re working in.

It’s challenging to maintain high-quality video and audio that people have come to expect from their everyday applications. When people are busy, the last thing they want to do is troubleshoot laggy video or figure out why they can’t hear a particular person. In Multi, the heart of every pairing session is the code, designs, or other content people are working on. Screen sharing and audio have to be reliably crisp, even in adverse network conditions. That’s why, from the very beginning, the team at Multi knew that they would build on Zoom Video SDK.

The team at Multi had prior experience building audio and video products in the gaming and productivity space, having used several home-grown and third-party infrastructure solutions. Home-grown solutions can work when the connections are localized or peer-to-peer, but the complexity is harder to manage once servers are required to scale for more participants. Zoom’s extensive and world-class server infrastructure helps to address the complexity of deciding which server should host a session.

They ruled out other providers due to two technical requirements: First, the audio and video quality wasn’t reliable enough, especially when it came to screen sharing. Second, they needed a robust SDK for native development on macOS. Zoom solved both of those problems. Of course, then there was the added bonus one gets when they build with Zoom Video SDK: when customers hear your app runs on Zoom, it helps instill confidence in your product.

Multi leans on Zoom for as much heavy lifting as possible. Although that started with real-time audio and video, they’ve since expanded to other features, including chat, transcription, and recording. Each of these features was able to launch with less than a week of work. That gives the team more time to focus on building out differentiated functionality, such as multicursor control or the ultra-low-latency video stream that’s used when the control is active.

“We’re building Multi to be the best way to get work done together. There are two parts to this: productive working sessions, and strong follow-through. Sync and async.” – Alexander Embiricos, Co-Founder and CEO

So far, that means focusing on features for multiplayer collaboration during sessions. Next, that will mean shifting to assisting with follow-up work, sharing context with teammates, and keeping team knowledge fresh. For example, they recently shipped summaries that include deep-linked reference code and designs that were shared. In the future, they expect to launch an AI assistant that shows those notes and helps automate the work when a person revisits that code or designs.

To experience the Zoom Video SDK in action, try Multi today.