Miguel Ángel
2.0 KiB
Narration Script
Video: HyperFrames Native Renderer — Design Proposal
Duration target: ~60 seconds Audience: HeyGen / HyperFrames engineering and product team Tone: Confident, direct, technical — peer to peer, not pitch
Eighty-six percent.
That's how much of render time Chrome spends just taking screenshots.
The native renderer changes that. H T M L stays the authoring format. A Rust and Skia I R replaces the per-frame browser path for the subset we control.
The auto backend chooses transparently — native when the compiler can prove support, Chrome when it can't. Every decision is logged. Nothing changes silently.
Adapters compile known primitives into a deterministic I R. G-SAP stays canonical. Text runs through Chrome once, cached as glyph clips, then composited.
Seven phases. Starting with proof infrastructure — not speed claims. Can we detect support accurately? Can we prove quality gates? Can the auto backend be trusted?
Then we grow the subset. One fixture at a time.
Native rendering isn't replacing Chrome. It's giving the best HyperFrames work a faster lane.
Let's build it right.
Word count: ~143 words (~57 seconds at 2.5 wps)
Beat cues
| Line | Beat |
|---|---|
| "Eighty-six percent." | Beat 1 — Hook |
| "That's how much of render time..." | Beat 2 — The Bottleneck |
| "The native renderer changes that..." | Beat 3 — The Solution |
| "The auto backend chooses..." | Beat 4 — The Contract |
| "Adapters compile known primitives..." | Beat 4 continued |
| "Seven phases. Starting with proof..." | Beat 5 — The Plan |
| "Then we grow the subset..." | Beat 5 continued |
| "Native rendering isn't replacing Chrome..." | Beat 6 — Closer |
| "Let's build it right." | Beat 6 — Closer |