LL-HLS (Low-Latency HLS) is Apple's standardized extension that brings HLS live latency down from 20-30 seconds to roughly 2-5. It works by publishing partial segments (small chunks of a segment) before the full segment is ready, plus playlist hints and blocking requests so the player gets new media almost as soon as it exists.
Because it stays within the HLS ecosystem, LL-HLS plays on Apple devices natively and reuses ordinary CDN infrastructure, though the CDN must support the required HTTP behaviors. The trade-off is the universal one for low latency: smaller buffers leave less cushion for network jitter, so resilience tuning matters more.


The handbook · 7 volumes
Fora Soft Video Engineering Handbook
LL-HLS brings HLS live latency from 20-30 seconds down to about 2-5 using partial segments and blocking playlist requests. The seven volumes cover the video stack; for LL-HLS start with Volume 2 on HLS, partial segments, and latency budgets, then Volume 1 on encoder settings, with CDN and player behavior in Volume 3. Written by Nikolay Sapunov, CEO at Fora Soft.
Rolling out LL-HLS and seeing stalls when latency drops?
We will test your packager, CDN, and players and send an LL-HLS plan with part duration and buffer settings. 250+ video products shipped since 2005.