DroiClaw Wants to Turn Smartphones Into AI-Native Devices

Artificial intelligence is moving beyond being a feature inside individual apps. The next stage could see AI become part of the operating system itself, allowing devices to understand what users want and complete tasks with far less manual input.

That is the idea behind DroiClaw (Zhuge), an Agentic OS ecosystem developed by Shanghai Droi Technology. The company is building an app-free approach to computing that combines on-device AI with cloud-based models, while focusing on security, user control, and visibility into how AI agents perform tasks.

From AI features to an AI-first operating system

Most smartphones today still follow a familiar model: users open an app, select a feature, and complete a task themselves. AI has been added to these systems through tools for writing, image editing, voice commands, and other functions.

DroiClaw takes a different approach by placing AI deeper into the operating system’s architecture.

Instead of treating AI as another application, the platform integrates it into areas such as system scheduling, interaction, task execution, and resource management. The goal is to make AI a fundamental part of how the device operates.

DroiClaw also supports multimodal interaction, allowing users to communicate through voice, text, images, files, and video. The system is designed to adapt to users’ habits over time.

Moving beyond the app-based smartphone

The company sees an opportunity to reduce the number of steps required for everyday digital tasks.

Instead of opening several apps to book a trip, organize information, or summarize documents, users can give an AI agent a natural-language instruction. The agent can then understand the request, determine the necessary steps, coordinate system functions and compatible services, and complete approved actions.

Sensitive operations can still require user authorization, while routine activities can be handled automatically.

This creates a different philosophy for operating systems. Traditional platforms often start by asking which app a user wants to open. An AI-native system instead starts with the user’s intended outcome.

Combining cloud AI with on-device processing

DroiClaw uses a hybrid AI architecture that combines cloud-based large language models with on-device AI.

Under normal conditions, cloud models handle more advanced tasks, including multimodal interactions, AIGC and content generation, complex information processing, and skill execution.

If network access becomes unavailable or cloud service credits have been exhausted, the system can automatically switch to an on-device model. That model is primarily intended to handle basic question-and-answer functions.

DroiClaw also supports multiple cloud AI models, including customized and privately deployed models, giving users and partners more flexibility over which AI services they use.

Security and user control remain central

As AI agents gain the ability to perform actions on behalf of users, control and transparency become increasingly important.

DroiClaw is designed around user authorization for sensitive tasks, along with permission management, task scheduling, and layered data protection. The platform also emphasizes operational visibility, allowing users and collaborators to better understand how tasks are being executed.

Local processing can also be used in situations where privacy is a priority.

The broader goal is to create an AI agent environment where automation does not mean giving up control.

Bringing AI to more affordable devices

Shanghai Droi Technology is also positioning DroiClaw as an accessible AI platform rather than something limited to expensive flagship smartphones.

The first smartphones using DroiClaw currently include devices from Coolpad and Philips, with prices of approximately RMB 1,099, or around US$160.

The platform is not designed to depend on a single smartphone manufacturer or AI model. It can support multiple device makers, developers, and AI providers.

Users can personalize AI avatars, install skills, create their own skills and scheduled tasks, and configure different AI models.

Built on years of OS experience

DroiClaw is backed by Shanghai Droi Technology’s more than decade-long experience in operating-system development. The company was founded in 2008.

According to Shanghai Droi Technology, its earlier FreemeOS platform has been deployed on more than 200 million devices globally and supported by more than 1,000 hardware and software partners across Europe, India, Southeast Asia, Latin America, and other regions.

As AI becomes increasingly integrated into everyday computing, the operating system could play a much larger role in determining how people interact with intelligent tools.

DroiClaw’s approach brings together system-level AI integration, an app-free Agentic OS vision, hybrid cloud and on-device intelligence, and an open ecosystem designed to expand access to AI.

Whether this model becomes the future of smart terminals remains to be seen, but it highlights a broader shift in the technology industry: the smartphone experience may increasingly revolve not around which app users open, but around what they want their device to accomplish.

Also read : Klaviyo Bug May Have Exposed New Users’ Passwords to Advertisers