AgiBot Convenience Store 24-Hour Livestream Concludes: 6 Genie G2 Robots Complete 1,730 Tasks
Six AgiBot Genie G2 humanoid robots worked continuously for a full day at a 24-hour convenience store in Xuhui, Shanghai, completing a cumulative total of 1,730 tasks, including greeting customers 683 times, retrieving and delivering 369 items, making coffee 147 times, providing breakfast service 126 times, restocking and inspections 277 times, and cleaning tables 128 times. The livestream attracted more than 2 million viewers across all platforms. The six robots served respectively as greeters, baristas, stockers, cashiers, and cleaners, and were centrally scheduled by Genie Master, an embodied Agent framework previously launched by AgiBot. Through the Embodied Execution Protocol (EEP), it connects task judgment, action execution, and result feedback, and stores operational trajectory experience in long-term memory.
Why read this
AI score 60
Six AgiBot humanoid robots worked continuously for a full day in a convenience store, disclosing the task volume and how they coordinated with the Genie Master scheduling framework.
Topics
Body · AI translation
Related topics top start Related topics top end Futures promotion begin Futures promotion end Miaopai begin Miaopai end Video player start Video player end Market chart begin Market chart end
Sina Tech News, November 4 morning news: Recently, the results of the Zhiyuan Robot Zhiyuan AI WAVE series underwent real-world testing. Six Zhiyuan Jingling G2 humanoid robots, in the roles of greeter (Ah Hao), barista (Ka Shao), stocker (Huo Bao and Juan Wang), cashier (Ding Jiang), and cleaner (Leng Jie), entered a 24-hour convenience store in Xuhui, Shanghai that was open for business and worked continuously for a full day—taking orders, making products, restocking, and cleaning—starting round-the-clock continuous operations and accepting online supervision from viewers across the internet.
The final results of the livestream challenge showed that the six Jingling G2 robots completed a total of 1,730 tasks, including receiving customers 683 times, retrieving and delivering 369 items, making coffee 147 times, providing breakfast service 126 times, restocking and inspection 277 times, and cleaning tabletops 128 times. The livestream attracted more than 2 million viewers online across the internet.
During the real-world testing, the six "new employees" had interlocking roles, covering the complete service process of a convenience store from opening and welcoming customers to daily cleaning. The two "examiners," experience officers Xiaoyang biubiu and Xiao Xiao Nathan, specially posed a variety of challenges from the customer's perspective: roadblocks, items placed out of order, voice-described tasks, and other test questions, completing two rounds of patrol tests.
It is reported that Zhiyuan's previously launched embodied Agent framework Genie Master serves as the shared Agent "brain" for the six robots, responsible both for scheduling tasks and connecting execution results. Through the Embodied Execution Protocol EEP (Embodied Execution Protocol), Genie Master connects task judgment, action execution, and result feedback. The Agent organizes tasks according to the robots' capabilities and on-site conditions; action strategies complete operations in combination with real-time observations, and execution results are returned to the Agent. Genie Master summarizes successful and failed work trajectories and stores the relevant experience in long-term memory to provide a basis for subsequent judgment and exception handling. news_keyword_pub,stock,sz300024 iframe below main text iframe below main text General statement General statement QR code at end start QR code at end start
QR code at end start QR code at end start Editor name and work code
Editor name and work code