Purpose Interhuman AI builds social intelligence infrastructure that enables any AI system to read, interpret, and respond to human behavior in real time. It is positioned as a developer-facing API platform that helps build products in conversational scenarios that can react to what is actually happening.
Main capabilities The platform provides the Inter-2 and Inter-2 Audio models, designed specifically for social intelligence. Unlike text transcription, which only records content, Interhuman processes speech, facial expressions, and body language from a single video stream, capturing "how" a conversation happens. The models can detect 12 social signals from video, audio, and text, and process all three together, aligned in time: send video, receive detected signals. Its models are built on behavioral science research into how people actually communicate, using observable patterns in voice, facial expressions, and body language to understand social behavior.
How to use Official materials show that developers can use this capability by sending video and receiving detected signals, but the specific API calling method, interface documentation, installation commands, and so on are not yet provided. Example application forms mentioned in the materials include: a Google Meet plugin (presenting social signal insights in real time during video meetings), an AI-moderated interview research tool (respondents answer on camera, and the system reads confidence, hesitation, and engagement together with the transcript), an AI avatar decision coaching tool, and a tool that can join any meeting and generate a report (explaining where speakers are confident, hesitant, or losing the audience's attention).
Applicable scenarios The directions listed in the materials include: sales and customer conversations, learning and training, research and interviews, meetings and collaboration, robotics and physical AI, and emerging applications. Existing customer cases involve evaluating whether process execution conforms to human-like expression, and transforming mock interviews from content practice into complete expression coaching.
Limitations At present, the materials do not provide information such as pricing methods, prices, platform support lists, specific onboarding steps, or quality and effectiveness guarantees; related content is not yet available.
