LiveKit Agents Examples

A collection of runnable Python demos and reference applications built with LiveKit Agents. The repository currently contains 104 examples that cover everything from single-file quickstarts to production-style, multi-agent systems with dedicated frontends.

What's Inside

• Voice, video, and telephony agents that demonstrate LiveKit's real-time APIs and the livekit-agents Python SDK
• Metadata-rich examples – every script now starts with YAML frontmatter (title, category, tags, difficulty, description, demonstrates) so tooling and LLMs can reason about the catalog
• A centralized index (docs/index.yaml) that lists every example along with its metadata for fast discovery and automation
• Complex demos that showcase advanced patterns such as multi-agent orchestration, RPC integrations, hardware bridges, benchmarking, and testing utilities

Discover the Catalog

• Browse docs/index.yaml for the complete list of examples, their descriptions, tags, and demonstrated concepts
• Use the frontmatter at the top of each script (inside a triple-quoted string) to inspect metadata directly in the file
• Many larger demos include their own README.md with architectural details or frontend instructions

Getting Started

Prerequisites

• Python 3.10 or newer
• pip (or another Python package manager)
• LiveKit account credentials (LIVEKIT_URL, LIVEKIT_API_KEY, LIVEKIT_API_SECRET)
• API keys for the providers you want to exercise (OpenAI, Deepgram, Cartesia, Anthropic, etc.)
• Node.js 18+ and pnpm (only required for demos that ship with a web frontend)

Installation

1. Clone the repository:

   git clone https://github.com/livekit-examples/python-agents-examples.git
   cd python-agents-examples

2. Create and activate a virtual environment:

   python -m venv venv
   source venv/bin/activate  # Windows: venv\\Scripts\\activate

3. Install Python dependencies:

   pip install -r requirements.txt

Configure Environment Variables

Create a .env file at the repository root with your credentials. At a minimum:

LIVEKIT_URL=your_livekit_url
LIVEKIT_API_KEY=your_api_key
LIVEKIT_API_SECRET=your_api_secret

Add provider-specific keys (OpenAI, Deepgram, ElevenLabs, Cartesia, etc.) depending on the examples you plan to run. Many demos read these values via dotenv.

Running Examples

Console-based demos

Most single-file examples can be launched directly via the CLI helper shipped with livekit-agents:

python basics/listen_and_respond.py console

The console argument opens an interactive session where you can speak (or type) with the agent. Swap in any other example path as needed.

Demos with companion frontends

Some complex agents include a frontend (look for directories such as *-frontend/). The typical flow is:

python complex-agents/medical_office_triage/triage.py dev  # starts the backend worker
cd complex-agents/medical_office_triage/medical-office-frontend
pnpm install
pnpm dev

Each project-level README documents the exact commands, required env vars, and how the backend and frontend communicate.

Tips for Exploring

• Use rg "---" -g"*.py" to quickly find frontmatter blocks or discover all scripts in a given category
• Many demos demonstrate interchangeable components (LLM, STT, TTS, VAD). Adjust provider classes or configuration to experiment with different vendors
• The metrics/, testing/, and benchmarking/ directories provide utilities for measuring latency, load, and agent quality
• docs/index.yaml can be consumed by tooling or LLMs to generate curated playlists, search experiences, or documentation

Additional Resources

• LiveKit Agents documentation: https://docs.livekit.io/agents/
• LiveKit Agents GitHub repository: https://github.com/livekit/agents
• Join the LiveKit community on Discord: https://livekit.io/community