We’re designing a large-scale WebRTC-based video platform and I’d love some input from people who’ve tackled similar problems.
Use case
- Live stream to 1,000–10,000 concurrent viewers
- At specific moments, move participants on the fly into thousands of 1-to-1 breakout rooms (pair discussions)
- After a few minutes, bring everyone back seamlessly to the main live stream
- All orchestration must be API-driven (rooms, users, transitions)
We’re currently evaluating video API / infrastructure providers (SFU + HLS, or hybrid approaches) that can realistically handle this scale with acceptable latency, reliability, and cost.
Questions for the community
- Which video APIs or platforms have you seen work reliably at this scale?
- Any strong opinions on managed APIs vs rolling your own SFU stack?
- Pitfalls to watch out for when switching thousands of users between broadcast and WebRTC rooms?
Side note
We’re also looking for an experienced WebRTC freelance developer (WebRTC + Node.js) to help design the architecture and build an initial PoC. If this kind of problem excites you and you have real-world experience with large-scale WebRTC systems, feel free to DM me.
Thanks in advance—happy to share findings back with the community once we’re further along.
Top comments (0)