Artificial intelligence is evolving from a feature that lives inside individual apps into something much more fundamental: the foundation of how devices operate. Instead of requiring users to open apps, move between menus, and complete tasks manually, next-generation AI platforms are being designed to understand what users want and act on their behalf with minimal intervention.
This shift is at the heart of DroiClaw, also known as Zhuge, developed by Shanghai Droi Technology. The platform aims to create an app-free Agentic OS ecosystem powered by a hybrid architecture that combines on-device and cloud-based AI. Its goal is to create a secure, controllable, and observable environment where intelligent agents can perform tasks while keeping users in control.
What Makes an Operating System AI-Native?
Many modern smartphones already include AI-powered features such as image editing, writing assistance, voice commands, and content generation. However, these tools are typically added on top of a traditional operating system that still revolves around individual applications.
DroiClaw takes a different approach by embedding AI into the underlying architecture of the operating system itself. Instead of treating AI as an isolated feature that users need to launch, the platform integrates intelligence into areas such as system scheduling, user interaction, task execution, and resource management.
For users, this could mean a more natural way of interacting with their devices. Rather than thinking about which application they need to open, they can focus on what they want to accomplish.
DroiClaw’s AI-centric architecture is designed to support a new generation of smart terminals. The platform supports multimodal interaction through voice, text, images, files, and video while also adapting to user habits over time.
From Apps to an Agentic OS
For decades, smartphones have been organized around applications. Whether someone wants to book a trip, summarize a document, send information, or organize files, the process often requires opening several apps and manually transferring information between them.
DroiClaw aims to move beyond this app-centric experience with an app-free Agentic OS ecosystem. Instead of simply providing access to applications, AI agents are designed to understand natural-language requests, determine the steps required to complete a task, coordinate system functions and compatible services, and carry out approved actions automatically.
The objective is not to remove user control, but to reduce unnecessary manual work. Sensitive actions can still require authorization, while routine tasks can be streamlined as AI agents coordinate activities across the operating system.
This represents a fundamental shift in how computing works. A traditional operating system asks users which application they want to use. An AI-native operating system focuses first on understanding what the user wants to accomplish.
A Hybrid Approach to AI
Delivering responsive AI requires a balance between performance, privacy, connectivity, and computing power. DroiClaw addresses these requirements through a hybrid architecture that combines on-device AI with cloud-based large language models.
Under normal conditions, cloud-based models provide the platform’s more advanced capabilities, including multimodal interaction, content generation, complex information processing, and AI skill execution.
When network connectivity is unavailable, or cloud service credits have been exhausted, DroiClaw can switch to an on-device model for basic queries. While the local model is primarily intended for simpler question-and-answer tasks, it allows essential AI functionality to remain available even when cloud services cannot be accessed.
The platform can also support multiple cloud-based AI models, including customized and privately deployed models. This gives users and organizations greater flexibility when selecting the AI services that fit their specific requirements.
Building a Secure and Controllable Agent Ecosystem
As AI agents become capable of performing increasingly complex actions, security and transparency become essential.
DroiClaw is designed around the idea that automated actions should remain manageable and visible to users. Authorization mechanisms are intended to protect sensitive operations, while permission management, task scheduling, and layered data protection help establish boundaries for AI agents.
The platform also emphasizes operational visibility, allowing users and collaborators to better understand how tasks are executed. Combined with local processing for appropriate privacy-sensitive scenarios, these features are intended to create a more trustworthy environment as AI becomes increasingly integrated into everyday computing.
The challenge for agentic operating systems is not simply making AI more capable. It is also making sure that users understand what their AI agents are doing and retain meaningful control over those actions.
Making Advanced AI More Accessible
Another important part of Shanghai Droi Technology’s vision is AI democratization. Rather than limiting advanced AI experiences to expensive flagship smartphones, the company is working toward making AI-native technology accessible across more affordable devices.
The first smartphones equipped with DroiClaw, including devices from Coolpad and Philips, are priced at approximately RMB 1,099, or around US$160.
The platform is also designed to be more open rather than being restricted to a single hardware manufacturer or AI model. Users can personalize AI avatars, install skills, create custom skills and scheduled tasks, and configure different AI models according to their needs.
This flexibility could allow the operating system to evolve alongside users as their AI requirements change.
Experience Meets the Next Generation of AI
DroiClaw builds on more than a decade of operating-system development by Shanghai Droi Technology, which was founded in 2008. According to the company, its earlier FreemeOS platform has been deployed on more than 200 million devices worldwide and supported by more than 1,000 hardware and software partners across Europe, India, Southeast Asia, Latin America, and other regions.
As the technology industry moves toward AI-native computing, the operating system is becoming increasingly important to how artificial intelligence is delivered and experienced.
DroiClaw represents one example of this transition, combining system-level AI integration with an app-free Agentic OS vision, hybrid on-device and cloud intelligence, and an open ecosystem designed to broaden access to advanced AI capabilities.
The bigger idea is that future smart devices may no longer be defined primarily by the apps installed on them. Instead, they could be defined by how effectively they understand user intent, coordinate services, and complete tasks on the user’s behalf.
If this vision becomes mainstream, the next generation of computing may look less like navigating a collection of apps and more like simply telling your device what you want and letting the operating system figure out how to make it happen.