The Furhat RDK comes with a suite of powerful tools to create new faces and gestures for Furhat, capture expressive movement from users, curate coordinated performances with Furhat's eyes, head movements, voice, and facial expressions, enabling developers and designers to make full use of Furhat’s expressive capabilities.

Users and attention

We've integrated high performance computer vision models for tracking and analysing multiple faces with Furhat's high definition wide-view camera. This comes with very accurate estimates of head pose and depth. We've also developed complex spatial and situation models that allow developers to easily handle multiple users stating and interacting with the robot.

Multi-sensory

Furhat comes with support for microphones, microphone arrays, and of course, cameras. The Furhat RDK also has high fidelity simulations of sensory input. The system comes with support for several speech recognition systems supporting up to 40 languages, and one of the world's largest collection of synthesised voices in up to 40 languages

Natural language understanding

Furhat's RDK is very open, flexible, and elastic. This means that developers can easily integrate external natural language processing and understanding tools. On top of that, Furhat comes with a machine learning powered Natural Language Understanding tool that has some of the world's most capable context modelling features, using intents, entities.