One node. 200 live streams.
Zero dropped frames.
When you go live to a dozen platforms at once, the question isn't "does it work" — it's "does it stay flawless when everyone else is doing the same thing." So we measured it.
The test
We pointed a single relay node at a growing wall of simultaneous live streams — 10, then 50, then 100, then 200 — each one ingested and relayed in real time, and watched for the first sign of strain: a dropped frame, a stutter, a stream falling behind.
The result
At 200 concurrent streams, every single one was still delivering in real time — 30 frames per second from start to finish, zero errors, nothing falling behind — and the node still wasn't maxed out.
Why it holds up: we don't re-encode
Most of the cost in a restreaming service is video encoding. We skip it. By default Stream Repeater stream-copies your feed — it passes your exact bytes through to every destination without decoding and re-encoding them.
What that means for you
Headroom. Your stream isn't crammed onto a busy box fighting for CPU — there's room to spare even under heavy load. And when we need more capacity, we add nodes rather than overload one. Reliability isn't a slogan here; it's something we measure and keep proving.