Born for GenAI・AI-native by design・Real-world service
Next-Generation AI Interactive Robots,
Built for Real-World Service
Born for GenAI・AI-native by design・Real-world service
Next-Generation AI Interactive Robots,Built for Real-World Service
40K+
Robots Sold
10+
Countries
1,000+
Schools &Families
Since 2024
Deployed at NTUH Hsin-Chu Branch
The Key Value of Building Physical AI Employees
Enable on-site service interactions to be scheduled, recorded, and analyzed – managed systematically like SaaS.
The Key Value of Building Physical AI Employees
Enable on-site service interactions to be scheduled, recorded, and analyzed – managed systematically like SaaS.
Three Core Capabilities
The ideal interaction endpoint for AI-powered services, enabling applications across diverse scenarios.
Three Core Capabilities
The ideal interaction endpoint for AI-powered services, enabling applications across diverse scenarios.
the Optimal AI Interaction Endpoint
More than an upgrade,
a full-scale leap in capability.
Specifications
the Optimal AI Interaction Endpoint
More than an upgrade,a full-scale leap in capability.
Specifications
FAQ
Kebbi 3 is an interactive service robot equipped with voice, screen, expression, and mobility capabilities. It serves as a physical AI interaction endpoint and is widely applicable in retail stores, pharmacies, medical institutions, elder care facilities, schools, kindergartens, hotel receptions, and smart factories
Yes, Kebbi 3 is specifically designed for AI interactions. It can integrate with Large Language Models (LLMs) and GenAI, connecting to NUWA Brain or cloud AI brains from partners (such as Vyin RIS) to become a smart shopping guide or consultation portal in real-world service environments.
Compared to the previous generation, NB3 features significant hardware upgrades:
Brain Core: It adopts the MediaTek MTK G520 processor, featuring 10 TOPS of Edge AI computing power for significantly faster response speeds.
Memory & Storage: Upgraded to LPDDR5 8GB RAM and eMMC 32GB ROM to ensure smooth operation of complex models.
Operating System: Upgraded to Android 15+, providing better compatibility and security.
Vision & Communication: Equipped with a 13-megapixel camera (FOV 81°) and supports Wi-Fi 6 and Bluetooth 5.3.
Expandability: Provides dual USB 3.0 OTG Type-C ports on the head and back, allowing easy connection to external sensors, infrared temperature cameras, or Ethernet.
Kebbi 3 is equipped with a 9100 mAh lithium battery. It takes about 2-3 hours to fully charge, and under normal usage (excluding continuous singing and dancing), it can operate for 7-8 hours.
Currently, it does not have waterproof or dustproof features. It is recommended to keep it away from water, dust, and other environmental factors that may cause damage.
Kebbi 3 is equipped with a 4-digital microphone array and supports AIVC smart noise cancellation, beamforming, and sound source localization. It can accurately filter out environmental noise and focus on human voices, and even supports voiceprint recognition and voiceprint noise cancellation to identify specific family or team members.
The default wake word is “Hello Kebbi”. General users cannot directly customize the wake word, but for enterprise customers with large-scale deployments, custom brand-specific wake words can be tailored through customized solutions.
Yes, while some pre-installed local applications can run without Wi-Fi, an internet connection is required to engage in voice conversations and fully experience the AI interactive features.
User conversations and data are backed up on AWS, the world’s leading cloud provider. Sensitive personal information is encrypted, and data will never be arbitrarily shared with third parties, complying with various international privacy regulations.
Yes. On the software side, NUWA offers three collaboration models:
Pure Version (Highest Customization): Provides a pure Android 15+ system and Robot SDK, allowing customers to build their own Apps, AI engines, and brand images entirely.
Customized Purpose (Semi-Customized): Adds or removes specific features based on the original factory functions.
Standard Purpose: Directly paired with NUWA RaaS Suite cloud tools for rapid deployment. On the hardware side, partial appearance (such as logo, color, ear accessories) and packaging customizations are available for a Minimum Order Quantity (MOQ) of 500 units or more.
FAQ
Kebbi 3 is an interactive service robot equipped with voice, screen, expression, and mobility capabilities. It serves as a physical AI interaction endpoint and is widely applicable in retail stores, pharmacies, medical institutions, elder care facilities, schools, kindergartens, hotel receptions, and smart factories
Yes, Kebbi 3 is specifically designed for AI interactions. It can integrate with Large Language Models (LLMs) and GenAI, connecting to NUWA Brain or cloud AI brains from partners (such as Vyin RIS) to become a smart shopping guide or consultation portal in real-world service environments.
Compared to the previous generation, NB3 features significant hardware upgrades:
Brain Core: It adopts the MediaTek MTK G520 processor, featuring 10 TOPS of Edge AI computing power for significantly faster response speeds.
Memory & Storage: Upgraded to LPDDR5 8GB RAM and eMMC 32GB ROM to ensure smooth operation of complex models.
Operating System: Upgraded to Android 15+, providing better compatibility and security.
Vision & Communication: Equipped with a 13-megapixel camera (FOV 81°) and supports Wi-Fi 6 and Bluetooth 5.3.
Expandability: Provides dual USB 3.0 OTG Type-C ports on the head and back, allowing easy connection to external sensors, infrared temperature cameras, or Ethernet.
Kebbi 3 is equipped with a 9100 mAh lithium battery. It takes about 2-3 hours to fully charge, and under normal usage (excluding continuous singing and dancing), it can operate for 7-8 hours.
Currently, it does not have waterproof or dustproof features. It is recommended to keep it away from water, dust, and other environmental factors that may cause damage.
Kebbi 3 is equipped with a 4-digital microphone array and supports AIVC smart noise cancellation, beamforming, and sound source localization. It can accurately filter out environmental noise and focus on human voices, and even supports voiceprint recognition and voiceprint noise cancellation to identify specific family or team members.
The default wake word is “Hello Kebbi”. General users cannot directly customize the wake word, but for enterprise customers with large-scale deployments, custom brand-specific wake words can be tailored through customized solutions.
Yes, while some pre-installed local applications can run without Wi-Fi, an internet connection is required to engage in voice conversations and fully experience the AI interactive features.
User conversations and data are backed up on AWS, the world’s leading cloud provider. Sensitive personal information is encrypted, and data will never be arbitrarily shared with third parties, complying with various international privacy regulations.
Yes. On the software side, NUWA offers three collaboration models:
Pure Version (Highest Customization): Provides a pure Android 15+ system and Robot SDK, allowing customers to build their own Apps, AI engines, and brand images entirely.
Customized Purpose (Semi-Customized): Adds or removes specific features based on the original factory functions.
Standard Purpose: Directly paired with NUWA RaaS Suite cloud tools for rapid deployment. On the hardware side, partial appearance (such as logo, color, ear accessories) and packaging customizations are available for a Minimum Order Quantity (MOQ) of 500 units or more.
