Privacy Policy

Local-First Architecture

DeskClaw is built on a local-first architecture. All core data processing — file operations, workflow execution, task management, and browsing automation — runs entirely on your local device. DeskClaw does not upload your files, task data, browsing history, or workflow results to NoDesk AI servers or any third-party cloud storage.

AI Model Communication

When DeskClaw processes tasks that require AI model inference, it communicates directly with the AI model provider you have configured (such as DeepSeek, Qwen, Kimi, GLM, or MiniMax). These API calls are made using your own API keys and go directly to the model provider's servers.

NoDesk AI does not proxy, intercept, or store the content of these API calls. The data handling policies of the model provider you choose apply to these communications.

Data You Control

The following data categories remain entirely on your local device:

• Files read, written, or organized by DeskClaw
• Browser automation session data and extracted content
• Task history, workflow configurations, and execution logs
• Skill configurations and custom automation settings
• Enterprise IM integration data (Feishu, DingTalk, WeCom messages and summaries)

Enterprise Edition

The DeskClaw Enterprise Edition is open-source, allowing organizations to audit the complete codebase and verify data handling practices. Organizations deploying the enterprise edition maintain full control over all data flows and can configure the system to meet their specific security and compliance requirements.

Updates to This Policy

This privacy policy may be updated as DeskClaw evolves. Significant changes will be communicated through the product's update channels. The date at the top of this page indicates the most recent revision.

Contact

For privacy-related inquiries, please contact NoDesk AI through the official channels listed on the DeskClaw product.