To set up a conversational AI web application using LiveKit, follow these steps: Prerequisites • A LiveKit Cloud Project or self-hosted LiveKit server • A Deepgram API Key for speech-to-text • An OpenAI API Key for LLM and text-to-speech • Python 3.9-3.12 Steps 1. Set up LiveKit account and install CLI: ◦ Create or sign in to your LiveKit Cloud account ◦ Install the LiveKit CLI and authenticate using

lk cloud auth

1. Bootstrap an agent from template: ◦ Clone the starter template using the command:

lk app create --template voice-pipeline-agent-python

◦ Enter your API keys when prompted ◦ Install dependencies and start your agent 1. Bootstrap a frontend from template: ◦ Create a frontend application to interact with your agent 1. Launch your app: ◦ Start your server with:

yarn dev

◦ Open

localhost:8000/room

in your browser Next Steps • Explore React Components for building more complex apps • Consider web and mobile frontends for your agent • Add telephony integration for call capabilities • Learn about deploying to production If you need further assistance, feel free to ask! Sources: AI Voice Assistant Quickstart | LiveKit Docs | Next.js 12 quickstart | LiveKit Docs | Voice AI quickstart | LiveKit Docs