📞 Inspiration

Customer inquiries, appointment confirmations, delivery logistics, and emergency escalations often require real-time telephone interaction. While text-based chatbots are ubiquitous, millions of real-world workflows happen over the phone.

We built VoiceAgentX: Autonomous Phone Dispatch & Voice Intelligence Assistant powered by the CALL-E SDK and Gemini / LLM reasoning to bridge the gap between autonomous AI agents and real phone telephony.


🚀 What It Does

VoiceAgentX acts as an autonomous outbound and inbound voice agent:

  1. Dynamic Call Planning: Formulates call goals, conversation outlines, and fallback rules before dialing.
  2. Real-time Voice Conversation: Holds low-latency, natural, context-aware spoken conversations with human responders over the phone.
  3. Structured Action Extraction: Converts phone call dialogues into structured CRM data, appointment bookings, and dispatch tickets.

🛠️ How We Built It

  • Telephony & Agent Interface: Built on CALL-E SDK & API for low-latency audio streaming and SIP call handling.
  • LLM Core: Fast conversational response generation with low latency streaming.
  • Backend: FastAPI webhook router with WebSocket audio pipes.

🏆 Accomplishments

  • Real-time conversational turn-taking with interruption handling.
  • Deterministic extraction of appointment times and customer intents.

Built With

Share this project:

Updates