Cekura (cekura.ai) is a San Francisco-based seed-stage company founded in 2024 that provides end-to-end testing and observability tooling specifically for conversational AI agents, including voice and chat. The platform lets teams run pre-production simulations against their agents using a curated library of caller personas with varied accents and background noise, as well as adversarial scenarios covering interruptions, long pauses, poor audio quality, and edge-case user behavior. In production, Cekura monitors live calls and tracks metrics such as latency, instruction-following accuracy, tool-call correctness, and CSAT scores. It sends real-time Slack alerts on quality drops or latency spikes, and provides call analytics for reviewing drop-offs, failed transfers, sentiment shifts, workflow adherence, and compliance failures. Cekura integrates natively with major voice-AI infrastructure providers including Retell, VAPI, ElevenLabs, LiveKit, Pipecat, and Bland, enabling teams to wire it into existing stacks with minimal setup. It also supports CI/CD pipeline integration for automated regression testing on every deployment. The platform is SOC 2 Type II certified, HIPAA compliant, and GDPR compliant, with features such as transcript redaction, role-based access control, and audit trails. A Model Context Protocol (MCP) server is available so AI coding assistants like Claude Code and Cursor can directly access Cekura's documentation and APIs. The company had raised $2.4M in seed funding as of 2025-2026. Key features: - Pre-production voice and chat agent simulation with diverse personas, accents, and background noise - Adversarial scenario testing covering interruptions, long pauses, VAD behavior, and WebRTC performance - Production call monitoring with latency, instruction-following, tool-call accuracy, and CSAT metrics - Real-time Slack alerts for latency spikes and quality drops in production - Native integrations with Retell, VAPI, ElevenLabs, LiveKit, Pipecat, and Bland - CI/CD pipeline support for automated regression testing on deployments - SOC 2 Type II, HIPAA, and GDPR compliance with transcript redaction and audit trails - MCP server for direct access by AI coding assistants
Free plan with 300 credits (no credit card required). Paid Developer plan and custom Enterprise plans available. Credits consumed at 5 credits/minute for voice simulations, 0.5 credits/reply for text simulations, and 0.2 credits/call for conversation analysis.