Skip to content

README.md

Latest commit

 

History

History
122 lines (76 loc) · 4.65 KB

README.md

File metadata and controls

122 lines (76 loc) · 4.65 KB
Raw

LiveKit Agents Examples

Looking for examples and guides? Most examples now live in the LiveKit docs. Browse the full collection of runnable examples and recipes on the Recipes page.

This directory contains various examples demonstrating different capabilities and use cases for LiveKit agents. Each example showcases specific features, integrations, or workflows that can be built with the LiveKit Agents framework.

Model Configuration

Most examples use LiveKit Inference by default for STT, LLM, and TTS models. This provides a unified API for accessing multiple model providers through LiveKit Cloud.

from livekit.agents import inference

session = AgentSession(
    stt=inference.STT("deepgram/nova-3"),
    llm=inference.LLM("google/gemma-4-31b-it"),  # low-latency gemma, hosted on LiveKit
    tts=inference.TTS("cartesia/sonic-3"),
)

Note: Realtime models (e.g., openai.realtime.RealtimeModel) are not supported by LiveKit Inference and must use the plugin directly. See the Real-time Models examples in voice_agents/.

📁 Example Categories

🏠 Homepage

A product knowledge agent demonstrating progressive disclosure with a Markdown-backed knowledge base, generated tool schemas, centralized prompt templates, session behaviors, streaming TTS filters, and a split unit/eval test suite.

🎙️ Voice Agents

A collection of voice-based agent examples, including basic voice interactions, tool integrations, RAG implementations, real-time models, and tracing.

🔄 Warm Transfer

Demonstrates supervisor escalation workflows for call centers, showing how to implement warm transfers where agents can brief supervisors before connecting them to customers.

🚗 Drive-Thru

A complete drive-thru ordering system example that showcases interactive voice agents for food ordering with database integration and order management.

🎭 Expressive Agent

A free-form companion agent running Expressive Mode. Tell it good news or bad news and hear the delivery change: emotion, pacing, and non-verbal sounds, all chosen from context.

🏢 Front Desk

A front desk agent example demonstrating how to build customer service agents with calendar integration and appointment management capabilities.

🔧 Primitives

Basic building blocks and fundamental examples showing core LiveKit concepts like room connections, participant management, and basic audio/video handling.

🛠️ Other

Additional examples including text-only agents, various TTS providers, transcription services, and translation utilities.

Running Examples

To run the examples, you'll need:

Environment file

Create a .env file in the examples directory and add your API keys (see examples/.env.example):

LIVEKIT_URL="wss://your-project.livekit.cloud"
LIVEKIT_API_KEY="your_api_key"
LIVEKIT_API_SECRET="your_api_secret"

When using LiveKit Inference (default for most examples), your LiveKit API key/secret is used for authentication. For examples that use provider plugins directly (e.g., realtime models), you'll also need the provider-specific API keys:

OPENAI_API_KEY="sk-xxx"  # For realtime models and provider-specific features
# ... other model provider API keys as needed

Install dependencies

From the repository root, run the following command:

uv sync --all-extras --dev

Running an individual example

Run an example agent:

uv run examples/voice_agents/basic_agent.py console

Your agent is now running in the console.

Taking an example with you

Each example directory stands on its own: copy it anywhere, and

uv sync
uv run agent.py console

installs from PyPI without the rest of the repo. lk agent deploy . works the same way, off the Dockerfile in the directory. Inside this repo the same directories are uv workspace members, so they run against your local SDK changes instead.

For frontend support, use the Agents playground or the starter apps.

📖 Additional Resources