navel research
Navel is an empathy robot for good care, made in Germany, designed for care facilities and homes as well as research institutions. It addresses each person individually by name in their native language, tailors its approach to each person's life topics and language skills, remembers previous conversations, chats about everyday topics, tells jokes and poems, maintains eye contact, and reacts empathetically. It is operated intuitively via natural language, configurable via a browser, works in any home with WLAN, and is data protection compliant (GDPR), with conversation data stored only on the robot and image data analyzed on-device and deleted after 100 milliseconds.
Primary use case: Social companionship and emotional support for people in need of care, improving well-being and quality of life in care facilities and homes; also available as a research platform with Python SDK.
Height
72 cm
Weight
8 kg
Specifications
Physical
Height
72 cm
Weight
8 kg
Locomotion
unknown
Manipulation
Hand Type
unknown
Power & Autonomy
Autonomy Level
unknown
AI & Compute
AI Capabilities
Face detection, Person identification, Emotion recognition, Head pose estimation, Eye gaze tracking, Sound source localization, Beam forming, Speech-to-text via cloud service (e.g. Microsoft), Text-to-speech with Acapela (30 languages), LLM-based dialog management (e.g. GPT), Personalized conversation addressing people by name in their native language, Memory of previous conversations, Continual learning of new skills through regular updates, Multilingual interaction in 74 languages, Socially aware proxemics, SLAM navigation (in planning)
Onboard Compute
NVIDIA Jetson AGX Xavier
Commercial
Availability
Purchase available for care (navel care) and research (navel research) variants; delivery time currently 10-14 weeks
Primary Use Case
Social companionship and emotional support for people in need of care, improving well-being and quality of life in care facilities and homes; also available as a research platform with Python SDK.
Related news
Asimov Releases Python SDK, Announces Open-Source Edge Control Layer
Asimov (Menlo Research) has released a Python SDK for its Asimov 1 humanoid under Apache 2.0, offering high-level walking commands, direct joint control, and access to the robot's camera, microphone and speaker via LiveKit or native UDP. The team also announced it is open-sourcing Asimov Edge, an arbiter that manages competing control sources and forwards commands to firmware, though the outlet could not locate the Edge source code at the time of review.
Asimov Open-Sources the Training Code Behind Its Humanoid’s Walk
Menlo Research's Asimov project has open-sourced the locomotion training code for its Asimov 1 humanoid, releasing an Isaac Lab-based framework with PPO and AMP configurations, reward definitions, actuator settings, and simulated variations for domain randomization. Although the announcement references an included trained policy checkpoint, Humanoids Daily's review of the linked GitHub repository and its releases did not find a downloadable checkpoint. The release complements Asimov's prior open-source CAD/simulation files and DIY kit shipments, giving developers a starting point to retrain locomotion policies for modified hardware builds.
Asimov 1 DIY Humanoid Kits Begin Shipping as Team Details Final Costs
Menlo Research and the Asimov team are shipping the first batch of their open-source Asimov 1 DIY humanoid kit, locking in a final delivered price of $20,000.