On April 21st, Xiaomi Company announced that Xiaomi miclaw, its AI agent product, officially launched a version suitable for PC, Mac and screen speakers, and simultaneously started the beta test. Users can go to Xiaomi community to apply for test qualification from now on. This expansion marks that miclaw has moved from the mobile phone to the desktop and home scenes, and the cross-device collaboration capability has achieved an important leap.

PC, Mac and speaker with screen: analysis of sealing and testing equipment

According to the official information released by Xiaomi, the equipment and system requirements supported by this test are as follows:

-PC side: need to run Windows 10 64-bit and above operating system;

-Mac side: it needs to run macOS 12 Monterey and above;

-Sound box with screen: support Xiaomi smart home screen 11.

With the addition of the above-mentioned new platform, Xiaomi miclaw has now covered many terminals such as mobile phones, tablets, PCs, Macs and screen speakers, forming a more complete cross-end use matrix. This also means that users can get a continuous and consistent AI agent service experience in different scenarios and different devices, further opening up the boundary between personal computing devices and home intelligent hardware.CIFF Shanghai believes that the extension of miclaw from mobile phone to desktop and family scenes provides a new entry point for human-computer interaction in smart homes.

Starting from mobile phone: the evolution road of Xiaomi AI agent

Xiaomi miclaw is an AI interactive test product built by Xiaomi based on the self-developed MiMo model. The product was launched for the first time on March 6 this year, and a small-scale sealing test was started. In the past more than a month, miclaw has accumulated a lot of user feedback and usage data on the mobile phone, laying a solid foundation for this multi-terminal expansion.

MiMo big model is an important technical layout of Xiaomi in the AI field, with core capabilities such as multimodal understanding, natural language interaction, task planning and execution. Different from the traditional voice assistant, miclaw, as an AI agent, can not only respond to user's instructions, but also independently understand complex tasks, disassemble and execute steps, and call different devices to cooperate. This "agent" mode is regarded as an important direction of the next generation of human-computer interaction.

PC and Mac: system-level execution, enabling desktop productivity.

On the PC and Mac side, miclaw shows strong system-level operation ability. Different from the light-weight interaction on the mobile phone, desktop users often face more complicated requirements of file processing, data sorting and multi-tasking. In view of these scenarios, miclaw provides functions such as document sorting, data analysis and batch file processing, and can work seamlessly with miclaw on mobile phones across devices.

Users only need to send natural language instructions on the computer side, and miclaw can automatically dispatch mobile phones, smart homes and other devices to complete tasks together. For example, in the office scene, users can say to miclaw on the computer side, "Help me organize all PDF files on my desktop and rename them by date." Miclaw will identify files and perform batch operations by itself, greatly improving work efficiency.

The highlight is the ability to transfer files across devices. Taking the file transfer between mobile phone and PC as an example, users can say to the mobile phone version of miclaw: "Send the file about Xiaomi car on the computer." The mobile version of miclaw will directly call the PC to automatically find the target file and send it to the mobile phone. The whole process does not require users to manually operate the computer, which truly realizes the cross-end collaborative experience of "one-word scheduling".

Behind this system-level execution ability is miclaw's deep understanding of the underlying interface of the operating system and authority management. It can simulate manual operation to complete a series of complex instructions on the premise of ensuring users' privacy and safety, thus upgrading AI from "chat assistant" to "digital executor".CIFF Shanghai believes that miclaw's system-level operation ability on the PC side is expected to improve the intelligent level of home office scenes, which is instructive for the integration of smart home and smart office.

Screen speaker version: for family scenes, you can use it without a mobile phone.

For multi-member family scenes, miclaw with screen speakers provides a new way of interaction. Users don't need to rely on mobile phones, they can use them by directly giving instructions to the speakers. This version supports voice wake-up and multiple rounds of continuous conversations. It can respond quickly to light tasks such as setting alarm clocks and querying weather, and can independently plan and implement complex tasks such as travel planning and family activities.

Specific application scenarios include: intelligent timing reminder, family information broadcast, travel planning, etc. Miclaw can combine calendar, map, weather and other multi-source information to generate executable plans and provide them to users.

In addition, the screen speaker version also supports multi-device linkage and automatic process creation triggered by natural language, without the need for users to manually configure complex automation rules. For example, the user only needs to say "I'm out", and miclaw will automatically turn off the lights at home, start the security mode, and simultaneously generate a to-do list on the mobile phone. The speaker is responsible for voice interaction, while the mobile phone and PC are responsible for in-depth operations such as subsequent editing and confirmation, and the three work together to complete a complete closed loop from instruction to execution.

This ability is especially practical for users who have a large number of Xiaomi smart home devices. In the past, the scene linkage that users need to manually set in Mijia App can be completed in just one sentence, which greatly reduces the use threshold of smart home.

Cross-end Collaboration: From "Device Interconnection" to "Intelligent Collaboration"

With miclaw covering more terminals, Xiaomi is building a cross-end collaborative ecosystem with AI agents as the core. Different from the simple device interconnection in the traditional sense (such as screen projection and file transfer), miclaw emphasizes "intelligent collaboration"-AI independently selects the right device and calls the right ability to complete the task according to the user's intention.

The launch of Xiaomi miclaw multi-terminal version not only enriches Xiaomi's AI product matrix, but also provides a new reference path for the development of domestic AI agents. At present, global technology giants are exploring the landing form of AI agents, and Xiaomi has a unique competitive advantage with its huge intelligent hardware ecology and cross-device system capabilities. From mobile phones to tablets, from PCs to speakers, miclaw is becoming an "intelligent hub" connecting people and devices. In the future, with the access of more third-party services and the continuous iteration of AI capabilities, miclaw is expected to play a key role in more scenarios such as office, home and travel.