Agora has announced an expanded partnership with OpenAI to deliver advanced conversational AI capabilities through its platform. By integrating OpenAI’s Realtime API into its Conversational AI Engine, the company aims to enable developers and enterprises to build human-like AI agents that operate seamlessly across voice, text, and other inputs in real time.
The partnership marks the first time a multimodal large language model has been embedded within the Agora platform. This step comes as more businesses seek natural, real-time engagement with customers, employees, and machines across industries such as robotics, automation, and customer service.
“Real-time multimodal interaction is the missing piece for AI agents to feel truly human,” said Tony Zhao, CEO of Agora. “By integrating OpenAI’s Realtime API into our Conversational AI Engine, we’re giving developers the tools to build experiences that are faster, smarter, and more natural than ever before.”
Features designed for natural interaction
The updated Conversational AI Engine introduces several new features to improve engagement with AI-powered agents. These include automated greetings for more natural onboarding, mixed-modality interaction that allows switching between voice and text within a session, and flexible turn-detection controls to fine-tune conversational flow.
Agora has also introduced Selective Attention Locking, a proprietary technology that filters out background noise and interruptions, ensuring smoother interactions. Together with its global real-time network infrastructure and developer toolkit, the platform is designed to speed up development, simplify integration, and enhance user experience.
Industry adoption and recognition
Robotics startup Carbon Origins has already adopted the new technology, integrating it into its Constellation AI solution to support hands-free operation of autonomous robots. The system automates complex checklists and system operations, helping operators concentrate on higher-level tasks.
“The combination of OpenAI’s Realtime API and Agora’s conversational AI technology enable hands-free control of our autonomous robot fleet,” said Amogha Krishna Srirangarajan, CEO and Founder of Carbon Origins. “The technology powers the automation of complex checklists and system operations in our Constellation AI solution, allowing operators to focus on strategic tasks and orchestration instead of manual execution.”
In recognition of its efforts, Agora was recently awarded Best Communications API at the 2025 API World Conference. The accolade highlights the company’s role in driving more natural and responsive real-time communication between humans and AI agents.
A global platform for real-time engagement
Headquartered in Santa Clara, California, Agora provides developers with APIs that power real-time engagement across applications and devices, from video and voice to interactive live streaming and chat. The company serves more than 1,700 organisations worldwide, enabling use cases ranging from social media and live shopping to education and telehealth.
With its expanded collaboration with OpenAI, Agora is positioning itself to capture growing demand for conversational AI experiences that closely resemble human interaction.