Open dispatches from the team building the future of live production. Technology decisions, architecture deep dives, industry analysis, and the unfiltered build journal.
I've spent the last year asking one question: why does producing a professional live broadcast still require $50,000 in hardware, three software licenses, and a broadcast engineer? The answer is that nobody has tried to solve it from the browser up. Until now.
We evaluated every major streaming ingest platform. Mux was the developer favorite. We went with IVS. Here's the technical reasoning behind a decision that shaped our entire delivery architecture.
At 1080p30, the browser's software H.264 encoder starved the compositor of CPU cycles. Here's what we learned, how we mitigated it, and why server-side GPU compositing is the real fix.
Most platforms pick one delivery method and live with the trade-offs. We built three paths into the architecture from day one. Here's why each one exists and when viewers hit each path.
Our public roadmap — not a promise sheet with dates, but a map of the territory we intend to cover. Server-side GPU compositing, AI voice translation, native macOS, and virtual production.
Every broadcast control room has a preview monitor and a program monitor. We put both in a browser tab — with drag-to-resize, a collapsible sidebar, and a TAKE button that feels like hardware.
Per-show IVS channel provisioning, engagement slideout, watch page v4, co-producer popup controls, guest state machine expansion, canvas compositor stream persistence, and more.