ConvoStack

A real-time voice agent platform powered by Deepgram for speech-to-text and text-to-speech. Built with a FastAPI backend handling voice stream processing and a React frontend for live conversation UI. Features low-latency bidirectional audio streaming, intelligent turn-taking, and LLM-driven conversational responses.

Key Metrics

Full-Stack

Architecture

end-to-end

Deepgram

Voice Engine

real-time STT/TTS

Highlights

  • Deepgram-powered real-time speech-to-text and text-to-speech pipeline
  • Low-latency bidirectional audio streaming via WebSocket
  • LLM-driven conversational logic with intelligent turn-taking

Technologies

PythonFastAPIDeepgramReactTypeScriptDocker
Source Code