Skip to main content

Allow a first-party cookie so we can count page views and see which pages and campaigns bring people here. Cookie policy

VARTA

VARTA architecture

VARTA splits a voice agent into two halves. DAKSH does the slow, language-specific LLM work once, before any call, and writes it down as workflow data. The turn engine reads that data during the call with rules first and the LLM last, so most turns are answered in microseconds from pre-built audio.