Listen to this article
LiveKit raises $45 million in Series B funding to expand its developer-focused voice AI infrastructure. The company launches Agents 1.0, a framework for building both open-ended and closed loop voice applications, alongside multilingual models and upgraded telephony tools. With scalable cloud deployment and widespread adoption, LiveKit strengthens its position in powering real-time, voice-driven AI systems.
From Open Source Roots to Industry Backbone
LiveKit began as an open-source initiative during the pandemic, aimed at simplifying digital communication. In just three years, it has evolved into a foundational voice AI infrastructure used by over 100,000 developers. The platform now supports more than 3 billion calls annually through LiveKit Cloud, its ultra-low latency edge network. The technology is actively deployed in industries ranging from healthcare to automotive services, powering real-time, voice-driven interactions across diverse use cases.
Examples of adoption include Hello Patient, which developed a voice agent for managing hospital workflows; Salient, which applies voice AI in loan servicing; and Podium, which integrates AI agents into organizational functions such as sales, marketing, and scheduling.
$45 Million Boost: What the Series B Funding Fuels
LiveKit announced $45 million in Series B financing, bringing its total capital raised to $83 million. The funding round is led by Altimeter, with continued support from its Series A leaders Jamin and Brad. Joining this round is Mike Volpi with his first investment from Hanabi Capital.
The company plans to use this funding to expand its team and accelerate the development of its voice AI platform. The objective is to create an all-in-one infrastructure that enables developers to build AI agents capable of engaging in complex conversations across industries.
AI Agents Get Smarter with the Launch of Agents 1.0
Agents 1.0 introduces a redesigned framework that simplifies the creation of voice AI applications. Developers previously faced limitations when trying to implement closed loop agents, particularly for deterministic tasks like IVR replacement or business process automation. The earlier approach of embedding business logic into large language model prompts, supported by function tools, often failed due to the probabilistic nature of LLMs.
Agents 1.0 enables developers to build more reliable systems by supporting discrete subtasks through workflow orchestration. It differentiates between open-ended agents, like those used in ChatGPT Advanced Voice Mode or Speak for language learning, and closed loop agents that handle structured processes such as patient intake or loan qualification.
Breaking Language Barriers with Multilingual AI Models
LiveKit introduced a larger semantic turn detection model with multilingual support. The model supports 13 languages including Chinese, English, French, German, Italian, Japanese, Korean, Portuguese, Russian, Spanish, Turkish, Dutch, and Indonesian. It operates with low latency, running inference in under 25 milliseconds for a 100-token context on a CPU.
The model can handle mixed-language conversations, making it suitable for global applications. Turn detection remains one of the most challenging aspects of voice AI, and this upgrade improves accuracy and usability across varied linguistic environments.
Telephony Tech Gets a Voice AI Upgrade
LiveKit’s open-source SIP telephony stack has reached version 1.0. Development began in September 2023, shortly after the launch of ChatGPT Voice Mode. LiveKit anticipated rapid adoption of voice AI in telecom, which relies entirely on voice interfaces.
The SIP stack now includes:
- High-definition audio
- DTMF support
- Cold and warm call transfers
- Automatic noise cancellation
Thousands of concurrent calls are handled by this stack, including emergency communications. A quarter of 911 dispatch centers in the United States use LiveKit, contributing to saving lives weekly. The system’s reliability and performance reflect its growing role in critical infrastructure.
Recommended: Atomic Secures $10 Million To Power Seamless Financial Connectivity And Drive Account Primacy
Scaling Voice Agents with Cloud-First Infrastructure
LiveKit Cloud Agents is now in closed beta. This service allows developers to deploy their voice agents without managing complex infrastructure. Unlike stateless web applications, voice agents are stateful and require constant inference, contextual listening, and turn management.
Cloud Agents handles lifecycle tasks including:
- Elastic provisioning based on location
- Load balancing
- Health checking
- Context migration
- Transparent failover
The platform deploys containerized agent code across LiveKit’s global edge network, managing the full devops cycle such as version control, rollbacks, and logging. LiveKit has been using Cloud Agents internally to accelerate its own deployments.
Building for Developers, Backed by Community Energy
LiveKit continues to prioritize open-source development and community engagement. Developers have shaped the platform by contributing feedback, features, and real-world applications. This collaboration has informed updates like Agents 1.0, multilingual support, and scalable infrastructure.
The company’s progress is directly tied to the activity of its developer base. Contributions and deployments by external teams have fueled both innovation and adoption, solidifying LiveKit’s place in the ecosystem of voice-first technology.
Why LiveKit Becomes a Critical Player in the Voice AI Landscape
The recent funding, paired with key product launches, positions LiveKit to support a wide spectrum of voice AI deployments. Its infrastructure is used not only in consumer-facing applications, but also in sectors with high-stakes operational demands.
The industry’s pivot to voice interaction, accelerated by tools like GPT-4o, has moved from speculative to practical. LiveKit’s open tools, multilingual models, and scalable architecture address the technical challenges developers face in deploying real-time voice applications.
As voice interfaces gain traction, LiveKit’s stack enables faster deployment, more reliable performance, and broader accessibility, making it central to how voice AI is implemented across industries.
Please email us your feedback and news tips at hello(at)dailycompanynews.com
