On June 24, 2026, XiaoPai Technology hosted its 'Brand Strategy Upgrade Conference and New Product Launch Event' in Shenzhen, officially introducing the inaugural home intelligent agent terminal, the ClawStation series, alongside the dedicated home care terminal, KinSight. This move catapults the smart home industry swiftly into the era of visual AI agents. As an invited media outlet, Leigongke was present at the event.

Image credit: Leigongke

Let's delve into the products first: Besides the 'Industry Edition' terminal, KinSight, tailored for the caregiving sector, XiaoPai Technology has rolled out three ClawStation series offerings: the comprehensive flagship home AI terminal, ClawStation Pro; the ClawStation Standard Edition, which leans more heavily on cloud computing; and the ClawStation Explorer Edition, which 'empowers' the Aurora TV box through an Android app. Simply put, the ClawStation Pro serves as an 'AI brain' outfitted with a local model (7B) and local computing capabilities, providing 'perception' and 'decision-making' support for the smart home system. It leverages PicClaw and OpenClaw Skill to manage other smart home products.

Image credit: Leigongke

What particularly piques Leigongke's interest is the perception prowess of the ClawStation Pro. Typically, smart home systems rely on infrared sensors and millimeter-wave radar to ascertain 'a person's location within a room.' On the plus side, these human presence sensors eschew cameras, allaying concerns about user privacy breaches. Nevertheless, they fall short in accurately identifying the user's identity and can only track larger entities like the human body. In contrast, the ClawStation Pro adopts a visual sensor approach: it can connect four cameras and harness local models and computing power to conduct AI analysis of video footage on the device itself, thereby 'comprehending' the happenings within the home.

Image credit: Leigongke

This endows the ClawStation Pro with a plethora of novel capabilities that traditional smart homes cannot match, such as precisely discerning whether the individual in the room is an elderly family member or a child, and consequently adjusting the air conditioner temperature proactively. Upon detecting trash on the floor, the ClawStation Pro can also proactively dispatch instructions to the robot vacuum cleaner via PicClaw. Naturally, visual perception also poses fresh challenges. Home cameras entail privacy considerations, and users are more attuned to data storage, video access, and permission allocation. Hence, the ClawStation Pro underscored its local analysis and edge computing attributes at the launch event.

Image credit: Leigongke

Regrettably, the ClawStation series products are not yet available for purchase by the general public, so the actual recognition accuracy and agent capabilities are yet to be confirmed. However, what is intriguing is that at the launch event, XiaoPai Technology also declared that it would open up camera device interfaces to users, offering standard Skills such as device search, security authentication, audio and video acquisition, and device settings. These have already been listed on skill stores like ClawHub. Users can access their cameras through intelligent agents such as OpenClaw and HermesClaw to construct their own application scenarios.

Image credit: Leigongke

From the vantage point of smart home industry development, Leigongke posits that ClawStation's perception solution, employing 'cameras + AI' as 'sensors,' actually furnishes a novel direction worthy of industry discourse. After all, a functional smart home system necessitates proactive perception and sustained learning. Under current technological conditions, when it comes to 'proactive perception' capabilities, what sensor can rival cameras?