This project provides an interactive browser-based playground to experiment with OpenAI's Realtime API, leveraging the LiveKit Agents framework. It's designed for developers and researchers interested in exploring real-time AI interactions and building applications that utilize streaming LLM capabilities.

How It Works

The playground consists of a Python-based agent built on the LiveKit Python Agents framework, which interfaces with OpenAI's Realtime API. This agent is paired with a Next.js web frontend. The architecture allows for direct interaction with the streaming API through a user-friendly browser interface, enabling rapid prototyping and testing of real-time AI features.

Quick Start & Requirements

• Agent Setup: Navigate to /agent, copy .env.sample to .env, set LiveKit credentials, create and activate a Python virtual environment (python -m venv .venv, source .venv/bin/activate), install dependencies (pip install -r requirements.txt), and run (python main.py dev).
• Web Frontend Setup: Navigate to /web, copy .env.sample to .env.local, set LiveKit credentials, install dependencies (pnpm install), and run (pnpm dev).
• Prerequisites: Python 3.9+, pip, pnpm, LiveKit Cloud or self-hosted LiveKit server.
• Resources: Requires running both the agent and web frontend locally.

Highlighted Details

• Demonstrates OpenAI's Realtime API capabilities.
• Built on the LiveKit Agents framework for Python.
• Web frontend developed with Next.js.
• Facilitates direct browser-based experimentation with streaming AI.

Maintenance & Community

The project is associated with LiveKit. Further community or roadmap details are not explicitly provided in the README.

Licensing & Compatibility

• License: Apache 2.0.
• Compatibility: Permissive license suitable for commercial use and integration into closed-source projects.

Limitations & Caveats

The project requires both a Python agent and a Node.js frontend to be running concurrently. Specific OpenAI API keys and LiveKit credentials must be configured via environment variables.