Inworld, an AI research lab and inference provider, announced that it has acquired Ultravox, a platform developers use to build real-time voice agents. Members of the Ultravox team have joined Inworld and will continue developing the platform under the combined organization. The transaction consolidates voice AI research, model inference, and real-time conversational infrastructure in one place to support the next generation of consumer AI applications.
Complementary Technologies and Strategic Fit
Ultravox contributes expertise in speech understanding, reasoning, task completion, turn-taking, and interruption handling for live conversations. Inworld brings top-ranked speech models such as Realtime TTS-2 and Realtime STT-1, along with model serving and real-time inference infrastructure used by consumer-facing AI applications. By combining these capabilities, the two organizations aim to improve the full conversation, from understanding user requests to generating responses that fit the moment.
Immediate Upgrade to Realtime TTS-2
The first benefit from the acquisition is available immediately to Ultravox customers. Built-in Inworld voices on the Ultravox platform now run on Realtime TTS-2, a text-to-speech model ranked at the top of the Artificial Analysis leaderboard. The model supports more than 200 languages, delivers latency under 100 milliseconds, and accepts natural-language direction that lets developers control pacing, tone, and emotional delivery through inline instructions.
No Disruption for Existing Customers
Existing Ultravox agents continue to run as usual, and no migration or code changes are required as part of this announcement. The TTS-2 upgrade comes at no additional cost, and existing voice IDs remain functional. Customers that use voices from another provider will see no immediate change, while support is provided through Inworld and the original Ultravox team.
Leadership Perspectives on the Acquisition
Inworld CEO Kylan Gibbs said Ultravox has built a platform that developers rely on for real-time voice agents, and he welcomed the team to Inworld. He described the integration of Realtime TTS-2 into Inworld voices as the first benefit being delivered to customers. Zach Koch, co-founder and former CEO of Ultravox, said Inworld is the right home for the platform and the right team to take it further.
Customer and Ecosystem Response
Akhilesh Damaraju, CEO of Equal AI, said his company was already working with Inworld's voice technology before the acquisition and is excited to see Inworld's voices and Ultravox come together. Equal AI builds a personal AI for aspiring Indians that manages calls and speaks the user's language. Damaraju called the combination a natural next step for the voice experiences his company is building for customers.
Future Development Plans
The acquisition is intended to advance speech-to-speech experiences that bring speech understanding, reasoning, and expressive voice generation closer together. Inworld says it will share more about those capabilities, as well as further improvements for Ultravox customers, in the coming weeks. Members of the Ultravox team will continue developing the platform under the combined organization.
The acquisition unites two complementary layers of the real-time voice AI stack, strengthening Inworld's position as a provider of models and inference infrastructure for consumer AI developers. By retaining Ultravox's developer platform and immediately upgrading its voice capabilities, the deal responds to a growing need for responsive, natural voice agents. The combined organization is now positioned to accelerate the development of next-generation speech-to-speech applications.