Cekura

Cekura (formerly Vocera, YC-backed) automates quality assurance for conversational AI agents across their full lifecycle, from pre-production simulation and scenario testing through live production call monitoring. Teams building voice or chat agents on stacks like LiveKit, Pipecat, Vapi, Retell, or ElevenLabs can wire Cekura into CI/CD so every prompt or flow change gets regression-tested against realistic scenarios before shipping, instead of being caught by an unhappy customer. In production, it scores real calls against configurable rubrics and surfaces failure patterns in dashboards for support and product teams to act on.

Cekura is built for companies operating conversational AI at scale in regulated or high-stakes contexts such as healthcare, insurance and BFSI, logistics, recruitment, and retail, where a silently degrading voice agent is expensive to discover late. It fills a gap generic LLM-eval tools leave open: agent behavior evaluation that is voice- and telephony-aware, not just text-transcript scoring, giving teams continuous confidence that their deployed agents are still behaving the way they were designed to.