LiveKit

LiveKit started as an open-source WebRTC platform that handles the hard parts of real-time communication: routing audio and video between many participants, scaling to large rooms, and working across web and mobile. Developers use its SDKs to build video conferencing, livestreaming, telehealth, and interactive broadcast features, and can run it on their own servers or use the hosted LiveKit Cloud.

The company became prominent in voice AI. The Agents framework lets developers build agents that listen, think with a language model, and speak back, with low latency, handling details such as turn-taking and interruptions. It supports telephony so agents can answer phone calls, plus plugins for speech recognition, language, and voice providers. LiveKit powers voice features in well-known AI products, which helped establish it as a standard for real-time AI.

The core is open source under the Apache 2.0 license, and LiveKit Cloud has a free tier and usage-based pricing for connection minutes and bandwidth, with enterprise options. Building voice agents still requires attention to latency, cost, and testing. Alternatives include Daily and Pipecat for voice agents, Twilio, Agora, and Vonage for communications, and Zoom’s SDKs. LiveKit suits developers building real-time AI and video applications.