Skip to main content
VoiceGateway profiles voice agents and the infrastructure they run on. Attach one line to an agent you already run and per-modality spend lands in the dashboard. Point the same tool at your LiveKit deployment for SFU health and node capacity, and hand it a load generator’s output to judge the telephony path. The core is framework-neutral: import voicegateway pulls neither LiveKit nor Pipecat. attach() observes, passively, and is the only meter. guard() controls: fallback, rate limits, budgets.
Start with the quickstart if you have an agent running, or what is VoiceGateway if you want the problem statement first.