天眼查显示，叠纸游戏成立女娲仿生（上海）智能科技有限公司，注册资本1000万元。叠纸创始人姚润昊直接持股60%并担任实际控制人，法定代表人为钱钟琴。从经营范围来看，这家新公司的指向性相当明确，智能机器人研发与销售、服务消费机器人制造、玩具制造与销售以及人工智能硬件销售。
Qichacha shows Papergames established Nüwa Bionic (Shanghai) Intelligent Technology Co., Ltd. with registered capital of 10 million yuan. Papergames founder Yao Runhao directly holds 60% and serves as actual controller; the legal representative is Qian Zhongqin. From the business scope, the new company's direction is quite clear—intelligent robot R&D and sales, service consumer robot manufacturing, toy manufacturing and sales, and AI hardware sales.
女娲仿生的成立并非无迹可寻。今年3月，招聘平台上出现一批叠纸挂出的新岗位，AI机器人结构负责人、机器人系统工程师、机器人工业设计师，薪资带到30至60K乘16薪，落地上海杨浦。岗位描述不是做个周边那么简单。SOC/MCU选型、高密度多层PCB、狭小空间硬件堆叠、语音视觉触觉IMU电机驱动显示通信模组集成、EVT/DVT/PVT全阶段推进、对接板厂贴片厂模具厂，基本是把消费级智能硬件从0到量产的链路全铺开了。同期，还有个很原型期的岗位叫AI陪伴玩具原型专家，要求是拥抱不完美但能跑，用开源方案和AI工具快速搓验证机，熟ESP32、STM32、树莓派，能自己选型、画板、写嵌入式、做结构、跑3D打印。
Nüwa Bionic's establishment was not without trace. In March this year, a batch of new positions appeared on recruitment platforms posted by Papergames—AI robot structure lead, robot systems engineer, robot industrial designer—with salaries reaching 30–60K × 16 months, located in Shanghai Yangpu. The job descriptions are not as simple as making merchandise. SOC/MCU selection, high-density multilayer PCB, hardware stacking in narrow spaces, integration of voice, vision, touch, IMU, motor drive, display, and communication modules, full-stage EVT/DVT/PVT advancement, interfacing with board factories, SMT factories, and mold factories—basically laying out the entire consumer-grade smart hardware chain from 0 to mass production. In the same period, there was also a very prototype-stage position called AI companion toy prototype expert, requiring "embracing imperfection but being able to run," using open-source solutions and AI tools to quickly build validation machines, familiar with ESP32, STM32, Raspberry Pi, able to select components, draw boards, write embedded code, do structural work, and run 3D printing.
这种表述很像团队在找先把东西做出来的人，而不是等架构定型后再招执行。8月27日，上海叠纸互娱网络科技有限公司经营范围变更，新增AI软件开发、AI硬件销售、玩具制造。目前，叠纸官网现在还挂着嵌入式工程师岗位，方向是AI交互与智能设备开发，要求有消费级智能硬件量产经验，但没写具体产品名。综合来看，叠纸的意图已不难揣测，从游戏里和角色谈恋爱，走向把角色塞进实体设备。但产品形态仍是一个未知项。女娲仿生究竟要做桌面机器人、毛绒陪伴玩具，还是可动手办，暖暖、恋与等IP是否会以人格化交互的方式嵌入硬件，目前产品走到了验证机阶段还是供应链打样，这些均无明确答案。
This phrasing sounds like the team is looking for people who can "get something built first" rather than hiring executors after the architecture is finalized. On August 27, Shanghai Papergames Interactive Entertainment Network Technology's business scope changed, adding AI software development, AI hardware sales, and toy manufacturing. Currently, Papergames' official website still lists embedded engineer positions, direction AI interaction and smart device development, requiring consumer-grade smart hardware mass production experience, but no specific product name. Overall, Papergames' intent is not hard to infer—from "dating characters in games" toward "stuffing characters into physical devices." But product form remains unknown. Whether Nüwa Bionic will make desktop robots, plush companion toys, or articulated figures; whether IPs like Nikki and Love and Producer will be embedded into hardware via personalized interaction; whether the product has reached validation prototype stage or supply chain proofing—none of these have clear answers.
不过，招聘信息给出的方向足够清晰。叠纸想用AIGC把角色变成可对话的服务，从虚拟交互延伸到物理陪伴。当前，游戏公司手里握着设定、声线、玩家情感沉淀，缺的从来不是需求侧的动力，而是硬件工程能力、量产良率、售后体系，以及那个最难的命题，怎么让一个AI陪伴设备摆在家里不尴尬。而叠纸已经完成了成立独立公司、调整经营范围、搭建硬件团队，接下来要等的，不再是叠纸做不做的疑问，而是女娲仿生第一次交出的答卷，是玩具级AI陪伴，还是能跑语音、视觉、动作的真机器人。
However, the recruitment information gives a clear enough direction. Papergames wants to use AIGC to turn characters into conversational services, extending from virtual interaction to physical companionship. Currently, game companies hold character settings, voice lines, and player emotional attachment; what they lack has never been demand-side motivation but hardware engineering capability, mass production yield, after-sales systems, and the hardest proposition—how to make an AI companion device "not awkward when placed at home." Papergames has completed establishing an independent company, adjusting business scope, and building a hardware team. What comes next is no longer the question "will Papergames do it" but Nüwa Bionic's first answer—toy-grade AI companionship or a real robot capable of voice, vision, and motion.
这件事放在前面讨论的具身智能产业脉络里，有一个值得单独看的角度。前面讨论过米果智能migo的情感陪伴机器人、雨果在东京游乐园靠拥抱和跳舞吸引游客、Character.AI被肯塔基州起诉聊天机器人怂恿用户自残。这些案例都指向同一个问题，AI陪伴这个方向，需求是真实的，但产品形态、技术路径、伦理边界都还没有收敛。叠纸的优势在于，它手里有现成的IP、声线和玩家情感沉淀，这些东西是其他机器人公司没有的。暖暖、恋与这些IP的玩家，对角色本身有情感连接，如果这个连接能从屏幕延伸到实体设备，需求侧的动力是现成的。但游戏公司的短板也很明显，硬件工程能力、量产良率、售后体系，这些不是靠IP就能补上的。
Placed in the embodied intelligence industry thread discussed earlier, this matter has an angle worth separate attention. Earlier we discussed Miguo Intelligent's migo emotional companion robot, Hugo attracting tourists at a Tokyo amusement park through hugs and dancing, and Character.AI being sued by Kentucky over chatbots encouraging self-harm. These cases all point to the same question—in the AI companionship direction, demand is real, but product form, technical path, and ethical boundaries have not converged. Papergames' advantage is holding ready-made IPs, voice lines, and player emotional attachment—things other robot companies do not have. Players of Nikki and Love and Producer have emotional connections to the characters; if this connection can extend from screen to physical device, demand-side motivation is ready-made. But game companies' shortcomings are also obvious—hardware engineering capability, mass production yield, after-sales systems—these cannot be made up by IP alone.
前面讨论过小尺寸人形机器人进入家庭的趋势，也讨论过万元级人形机器人被买回去主要是玩而不是用。叠纸这个案例，走的是另一条路，不是先造一个通用机器人再找场景，而是先有IP和情感连接，再倒推需要什么样的硬件形态。这个逻辑和前面讨论的FDE逻辑、当虹科技的自主作业加远程接管、贝塔无限的具身长记忆，是同一个产业阶段的不同尝试，大家都在找具身智能真正能落地的具体形态。叠纸能不能把这个尝试做成，取决于两件事，一是硬件团队能不能把产品从验证机做到量产，二是玩家愿不愿意为一个实体化的角色买单。这两个问题的答案，要等女娲仿生的第一款产品出来才能知道。
Earlier we discussed the trend of small-size humanoid robots entering homes, and ten-thousand-yuan humanoid robots being bought mainly for play rather than use. Papergames' case takes another path—not building a general robot first and then finding scenarios, but having IPs and emotional connections first, then working backward to what hardware form is needed. This logic, alongside the previously discussed FDE logic, Wanghong Technology's autonomous operation plus remote takeover, and Beta Infinite's embodied long-term memory, are different attempts at the same industry stage—everyone is looking for the specific form where embodied intelligence can truly land. Whether Papergames can make this attempt work depends on two things: whether the hardware team can take the product from validation prototype to mass production, and whether players are willing to pay for a physicalized character. The answers to these two questions will only be known when Nüwa Bionic's first product emerges.
