Build realtime AI. Instantly transport audio and video between LLMs and your users.
Powering apps for
customers including:
customers including:
Solutions
Tools for multimodal apps
Voice AI
Robotics
Livestreaming
Build realtime multimodal experiences
Stream voice and video between your model and users with sub-100ms latency.
Simulate human conversation
Integrated plugins make speaking with AI as natural as speaking with a person.
Interruption handling
Turn-taking cue detection
Connect to telephony networks
Allow users to dial into sessions or make programmatic calls out to phones.
Supports DTMF
Call our demo voice bot:
developer focused
Build, deploy, scale. Repeat.
Realtime telemetry and analytics
Explore session and user data in depth to understand a user's experience and resolve issues.
Purpose-built components
Build realtime experiences quickly with higher-level abstractions for common patterns and use cases.
voice-agent.py
components.tsx
stream-ingest.go
mobile.swift
1async def entrypoint(ctx: JobContext):2 await ctx.connect(auto_subscribe=AutoSubscribe.AUDIO_ONLY)3 agent = VoicePipelineAgent(4 vad=silero.VAD.load(),5 stt=deepgram.STT(),6 llm=openai.LLM(),7 tts=openai.TTS(),8 allow_interruptions=True,9 )10 participant = await ctx.wait_for_participant()11 agent.start(ctx.room, participant)12 await agent.say('Hello, how can I help you today?')13
1
SDK support across all major platforms
Code your application with a LiveKit client natively designed for your platform of choice.
No vendor lock-in
Self-host LiveKit for free without changing a line of code. The entire ecosystem of tools and services is Apache 2.0 open source.
global scale
The backbone of the realtime computing era
LiveKit's network is optimized for ultra-low latency, extreme resiliency, and massive scale. Our team is distributed across the world and our infrastructure delivers billions of minutes of audio and video every month.
Uptime
99.99%
Latency
<100ms
capabilities
A feature-rich platform
SSO and RBAC for teams
Set up roles and collaborate on your realtime project with your team.
Enterprise-grade security
End-to-end encryption, SOC2 Type 2, GDPR, CCPA, and HIPAA compliant.
Noise and echo cancellation
Remove background artifacts from audio streams.
Session recording
Composite, record, and store sessions in any S3-compatible cloud bucket.
Stream ingestion
Import external video or audio streams from a myriad of formats.
Moderation tools
Restrict user access to publish or subscribe to any session or stream.