The all-in-one
Voice AI platform

Build, deploy, and scale realtime agents. Open source. Enterprise scale.

Open source voice agent framework
Ultra low-latency edge infrastructure
SOTA Voice AI tools and research
Powering billions of calls
in production for:
Open AI
Speak
Character AI
Hello Patient
Tinder
/logos/open-ai.svg
LiveKit Cloud powers ChatGPT’s Advanced Voice Mode for millions of users around the world, every day.
Run millions of concurrent calls
Automatic turn detection and interruption handling
Self-host or deploy agents to LiveKit Cloud
iPhone device for ChatGPT Advanced voice mode

Quickly build voice agents

Preview
agent.py
visualizer.tsx
1session = AgentSession(
2  stt=deepgram.STT(),
3  llm=openai.LLM(model="gpt-4o"),
4  tts=cartesia.TTS(),
5  vad=silero.VAD.load(),
6  turn_detection=turn_detector.EOUModel(),
7)
1
Voice AI quickstartBuild a simple voice agent with Python in less than 10 minutes.
Web and mobile frontend guidesBring your agent to life through a web or mobile app.
Integrate with telephonyEnable your voice agent to make or take phone calls.
Agents playgroundA virtual workbench to iterate on and test your voice AI agents.

How it works

1
User speaks to agent via app, browser, or phone call
2
User speech is streamed from device to agent via LiveKit Cloud
3
Agent receives user speech and runs your custom business logic
4
Agent generates and streams speech back to user via LiveKit Cloud
Step 1
Hello, I'm here.
Step 1
Step 1
Step 1
Built for scale

Enterprise grade
infrastructure

GDPR
HIPAA
SOC 2 Type 2
99.99%Uptime
100msGlobal latency
3 billionCalls annually
100K+Developers

Ready to build?

Start building your realtime application with a free account. Ping us if you're interested in custom pricing.

No credit card required • 50GB free monthly