but can be given verbally

A developer-first platform for building low-latency, conversational voice AI agents that handle complex tasks through natural speech.

Vapi orchestrates the full voice stack (STT, LLM, and TTS) to deliver sub-800ms response times for AI assistants. Developers use the platform to deploy agents that handle real-time interruptions and complex logic over phone lines or web interfaces. By integrating with providers like Deepgram for transcription and ElevenLabs for synthesis, Vapi enables businesses to automate 24/7 customer support and outbound scheduling with human-like verbal precision.

https://vapi.ai/