DroiClaw AI-Native OS Brings System-Level Agents to Affordable Smartphones at RMB 1,099

Shanghai Droi Technology Introduces DroiClaw as AI-Native Operating System

Shanghai Droi Technology has launched DroiClaw, an AI-native operating system that fundamentally reimagines how users interact with smart devices 1

1

. Unlike conventional operating systems where AI capabilities exist as separate applications or add-on features, DroiClaw integrates artificial intelligence directly into the underlying system architecture 3

3

. The architecture, interaction model, model scheduling, data processing, agent capabilities, and overall user experience are developed with AI as the foundation rather than as an afterthought 1

1

. This system-level integration allows AI to draw on data, coordinate tasks, and operate across the entire system rather than remain confined within isolated app silos 4

4

.

Moving Beyond Traditional App-Based Models Toward Agentic Era

DroiClaw's core objective is to create an app-free, agentic OS ecosystem that simplifies how people interact with their devices 1

1

. Instead of requiring users to open multiple apps and complete each step manually, the system understands user intent, breaks down requests, and determines the actions needed to complete a task 1

1

. Users can communicate naturally with DroiClaw through voice, keyboard input, images, documents, files, and real-time video queries 1

1

. This multimodal interaction capability enables the system to process different formats and coordinate multiple functions across supported device features and authorized third-party applications 3

3

. The agent-driven experience becomes increasingly personalized over time as the operating system adapts to user habits and preferences 1

1

.

Hybrid AI Architecture Combines On-Device AI Models and Cloud-Based Large Models

DroiClaw employs a hybrid AI architecture that combines lightweight on-device AI models with cloud-based large models to support different usage scenarios 1

1

. Under normal conditions, users primarily access cloud-based large models, which support advanced capabilities such as multimodal interactions, AIGC, complex reasoning, scheduled tasks, and skill execution 4

4

. On-device models serve mainly as fallbacks in specific situations, such as when network access is unavailable or cloud service credits have been exhausted 2

2

. In these cases, the system automatically switches to the local model to process basic queries without requiring manual intervention 4

4

. DroiClaw supports multiple cloud-based AI models, including customized and privately deployed options, providing users with flexibility in selecting AI services for different devices and use cases 1

1

.

System-Level AI Agents Enable Cross-System Task Coordination

What distinguishes DroiClaw from traditional chatbots is its ability to take action on behalf of users 2

2

. Users can employ natural language interactions to make requests like "Use Uber to book me a ride to JFK Airport," and DroiClaw performs the necessary operations 2

2

. The system can handle shopping, summarizing group chats, translating conversations, or scheduling recurring tasks 2

2

. Cross-system task coordination allows the operating system to interpret operations or requests and use apps on the user's behalf with proper permission 2

2

. Users can install skills from a skill library or create customized skills for specific tasks and workflows, allowing the system to adapt to different needs 3

3

. The system records frequently used skills and prioritizes them in future interactions, making skill invocation more convenient and relevant over time 4

4

.

Building Secure and Observable Agentic Ecosystems with User Control

As system-level AI agents become more capable of operating devices, accessing information, and coordinating services, the need for visibility and user control becomes critical 1

1

. DroiClaw aims to create a secure, controllable, and observable AI agent ecosystem by giving users greater oversight of how agents operate 1

1

. Sensitive operations require user authorization before execution, while task and model scheduling help determine how requests are processed 1

1

. The operational permission management system provides users with visibility into agent actions and when permissions are required 3

3

. For privacy-sensitive information, local processing provides an additional layer of protection 1

1

. This transparency helps build trust in agentic ecosystems while allowing users to maintain control over automated actions 3

3

.

Making Advanced AI Accessible Beyond Premium Devices

Shanghai Droi Technology, an established China-based mobile software and device integration company with roots in designing mobile phone motherboards, recognized that small and medium-sized smartphone manufacturers faced an AI need that couldn't be filled by standalone AI features 2

2

. DroiClaw aims to bring advanced AI capabilities to mainstream and lower-priced smart terminals rather than limiting them to premium flagship devices 1

1

. The first smartphones equipped with DroiClaw are priced at approximately RMB 1,099, reflecting the company's commitment to making AI-native operating systems accessible to broader user groups 1

1

. Through model flexibility, hardware neutrality, and ecosystem partnerships, the platform is designed to support a broader AI agent ecosystem across different smart devices 1

1

.