March 18, 2025
Artly AI, an AI robotics company specializing in service industry automation, has showcased its robotic hand upgrade for its Barista Bot at the Nvidia GPU Technology Conference 2025, according to a press release. This advancement aims to enhance the robot's dexterity, bringing it closer to human-level precision in crafting artisan coffee drinks.
Artly's Barista Bot, designed for use in coffee shops, event venues and restaurants, operates on the company's Vision-Language-Action AI model. This model is trained using motion capture data from human baristas, enabling the robot to replicate their movements and prepare 28 hot and iced beverages. The robotic hand upgrade improves upon the previous gripper system, allowing for better manipulation of everyday objects.
According to the press release, the upgrade addresses the limitations of customized grippers, which lack the range of motion and flexibility of human hands. Artly's robotic hand, featuring up to 20 degrees of freedom, force feedback and high-precision tactile sensors, allows the Barista Bot to handle various human-designed tools and objects, such as cups and spoons. Artly is also developing additional sensors, including temperature sensors, for food and beverage applications.
"With this new upgrade, Barista Bot now moves with even greater agility, mimicking human brewing techniques like never before," Yushan Chen, chief operating officer of Artly, said in the release. "By integrating our VLA with cutting-edge robotics, we ensure each cup is crafted with the skill and consistency of an award-winning barista. Our mission is to share culinary artistry with everyone, and we invite GTC attendees to experience it firsthand."
Artly's robotic platform is designed for versatility, with the capability to perform tasks beyond coffee-making, such as slicing fruit, managing glassware and preparing cocktails, making it suitable for restaurants and self-service kiosks in high-traffic locations. The company also notes that the robotic hand's alignment with motion-captured training data facilitates rapid training for applications in other sectors, including healthcare and logistics.