1 total
Both official SDKs shipped on 2026-10-16: voices created from a text description (Live only), 30-second Live audio fixtures, Decisions transcript triggers, and prompt caching for the Decisions API - all from the CHANGELOGs and tag sources.