On June 24, 2026, Xiaopai Technology officially released two core intelligent terminal products at the “Brand Strategy Upgrade Conference and New Species Launch Conference” held in Shenzhen: the first home intelligent terminal ClawStation series, and the dedicated home care terminal “KinSight”. This move also marks that the domestic smart home industry has officially entered a new stage of visual AI Agent development.
The products launched at this conference cover two major scenarios: consumer-level and industry-level scenarios: among them, “KinSight” is an industry version terminal specially built for the nursing industry, which can meet the centralized care needs of nursing homes and childcare institutions; the ClawStation series for ordinary home users has launched a total of 3 products with different positioning, including ClawStation, a full-featured flagship home AI terminal equipped with independent local computing power
Pro, the ClawStation standard version that relies on cloud computing power to lower the threshold of use, and the ClawStation exploratory version that can empower Aurora TV boxes and be implemented in the form of Android App, covering user groups with different consumption capabilities and different usage needs.
As the flagship of the series, ClawStation
Pro is the core highlight of this release. It is equipped with a large local model of 7B parameters and comes with an independent computing module, which is equivalent to the “AI brain” of the entire smart home system. It can not only complete the environment perception of the entire scene, but also output intelligent decisions that adapt to family needs. It can also achieve unified control of different brands of smart home products through the PicClaw and OpenClaw Skill ecology, solving the pain point of the inability to link different brands of smart homes in the past.
Different from traditional smart home sensing solutions that rely on infrared sensors and millimeter-wave radars, ClawStation Pro innovatively uses a visual sensor solution that can be connected to up to 4 home cameras. All video analysis work is completed on the device side through local computing power, without the need to upload to the cloud server. This not only circumvents the shortcomings of traditional sensors that cannot identify users and can only track large targets, but also eliminates users’ concerns about privacy leaks in visual solutions.
Relying on this capability, ClawStation Pro can realize many functions that cannot be accomplished by traditional smart homes: for example, it can accurately identify whether the person currently in the house is an elderly person or a child, and adjust the air conditioning temperature and light brightness in a targeted manner; when it recognizes that there is scattered garbage on the ground, it will actively send cleaning instructions to the sweeping robot; when it recognizes abnormal scenarios such as an elderly person falling down or a child staying alone for too long, it will push a reminder to the user as soon as possible.
At this conference, Xiaopai Technology also announced that it will open the camera device interface to users, providing standard skills such as device search, security certification, audio and video acquisition, and device settings. Related tools have been put on the shelves in skill stores such as ClawHub. Users can call their own camera devices through agents such as OpenClaw and HermesClaw to build their own smart home application scenarios. The playability and customization space have been greatly improved.
Currently, the ClawStation series of products are not available to ordinary consumers for purchase. The actual recognition accuracy and scene linkage fluency have yet to be verified by subsequent actual tests. However, the “local AI computing power + end-side visual perception” technology path adopted by this product has provided a new direction worth exploring for the smart home industry that currently encounters development bottlenecks.
Under the current technical conditions, visual sensors are home sensing devices with the richest sensory dimensions. With local computing power to solve privacy pain points, the implementation of smart homes that truly realize active services and autonomous learning will become closer and closer.
