What problem does it solve? Setting up a voice or chat AI agent on the Cekura platform involves many provider-specific steps — credentials, connection modes, mock tools, knowledge bases, and dynamic variables — and missing any of them produces broken test runs. This Skill walks through the entire setup as an 11-phase guided workflow and verifies the agent end-to-end with a real test conversation. ## Core Features & Use Cases - Provider-aware setup flow: Supports VAPI, Retell, ElevenLabs, LiveKit, Pipecat, Bland, Synthflow, KoreAI, Genesys, Cisco, and self-hosted agents, with per-provider credential and connection requirements. - Auto-import path: For VAPI, Retell, ElevenLabs, Bland, and Synthflow, imports the system prompt, phone number, tools, knowledge base, and dynamic variables automatically via configure_from_provider. - SDK integration: Wires the Cekura SDK into LiveKit or Pipecat agent code for transcripts, tool calls, metrics, and OTel traces. - Mandatory verification run: Generates a scenario, runs it against the agent, and inspects the transcript until a real back-and-forth conversation is confirmed. - Use Case: A developer with a Retell voice agent wants to start testing it on Cekura — the Skill collects the API key and agent ID, auto-imports the configuration, and confirms the agent works with a live test call. ## Quick Start Ask the assistant to create and connect your voice agent on Cekura, mentioning your provider such as VAPI or LiveKit.
