Xiaomi officially launched its mobile Agent interaction testing product "Xiaomi miclaw" today, marking a critical step forward in its smart terminal automation interaction field. The product is built based on Xiaomi's self-developed MiMo large model and is positioned as an AI agent application similar to "OpenClaw", aiming to deeply understand complex instructions through the large model and achieve environmental perception interaction.
As the first domestic application similar to OpenClaw to be implemented on a smartphone, Xiaomi miclaw currently adopts a high-entry threshold access mechanism, with the first batch of users being only Xiaomi 17 series device enthusiasts and tech fans, who will participate in a small-scale closed test through an invitation system.
This launch not only extends the capabilities of Xiaomi's MiMo large model but also demonstrates the underlying technical reserves of Pengpei OS in on-device AI deployment. By introducing an Agent (AI agent) into the mobile field, Xiaomi is trying to change the traditional human-computer interaction logic, shifting from "users searching for functions" to "AI proactively scheduling services."
Combined with the previously open-sourced 309 billion parameter MiMo-V2-Flash model and hardware layout such as AI glasses, Xiaomi is accelerating the construction of a cross-device AI ecosystem centered around proactive intelligence. In the context of homogenization in the smartphone market, the self-research and implementation capability of system-level Agents has become a key factor for top manufacturers to build a technological moat.
