DeskClaw is an AI-powered desktop agent built on the OpenClaw ecosystem. It lives on your screen, automates your browser, manages files, and orchestrates workflows — while keeping all data local and private.
DeskClaw is a desktop AI assistant that combines browser automation, local file operations, and multi-model intelligence into a single, always-available agent on your screen.
DeskClaw controls your browser natively. It can navigate web pages, extract data, fill forms, and execute multi-step web workflows — one of the most complex capabilities to set up manually with OpenClaw, built in from day one.
Read, write, and organize files on your local machine. DeskClaw accesses your file system directly to handle document processing, data extraction, and content creation without uploading anything to the cloud.
Connect to leading AI models including MiniMax, Kimi, GLM, Qwen, and DeepSeek. Choose the model that best fits your task, switch freely, and keep full control over your AI backend configuration.
All data processing runs on your local device. DeskClaw does not upload your files, browsing history, or task data to external servers. Your workflows and data remain entirely under your control.
Deeply adapted for Feishu (Lark), DingTalk, and WeCom (WeChat Work). Add DeskClaw to group chats as a bot that summarizes conversations, creates tasks, updates project documents, and posts to group announcements automatically.
Pre-configured skills for market research, information aggregation, workflow automation, and more. No setup needed — issue a command and DeskClaw executes end-to-end. Skills are extensible via the MCP (Model Context Protocol) ecosystem.
DeskClaw is designed for one-click installation. No command-line setup, no environment configuration, no API key juggling.
Download and install DeskClaw on your desktop. The entire OpenClaw runtime is packaged into a native application — no CLI, no manual dependency management. The setup process takes minutes, not hours.
Once installed, DeskClaw appears as a small animated crab on your desktop — an AI companion that stays with you across all applications. It is always available for commands, not a dialog window you open and close.
Tell DeskClaw what you need in plain language. It breaks down complex tasks into executable steps, controls your browser, reads and writes files, calls local applications, and delivers results — all while you watch the progress in real time.
As your needs grow, extend DeskClaw with additional skills. The Skill and MCP architecture lets you add specialized capabilities — from e-commerce analytics to content generation — as modular, reusable components.
DeskClaw replaces hours of manual work across research, operations, and communication tasks.
A founder exploring the cycling market used DeskClaw to batch-scrape user comments from YouTube cycling videos, automatically categorizing pain points and unmet needs. A task that previously required writing custom scrapers or manually reviewing hundreds of comments was completed with a single natural-language instruction.
An AI company founder configured DeskClaw to curate 20 of the most important AI industry news items each day, automatically aggregating, ranking, and generating summaries. The briefing is ready every morning, eliminating the need to manually scan information feeds.
Inside NoDesk AI, every team member has a DeskClaw agent connected to Feishu. The AI PMO bot "XiaoJinJu" automatically summarizes group chats, creates Feishu tasks, updates project documentation, and writes task summaries directly into group announcements when mentioned.
DeskClaw handles end-to-end information workflows: collecting data from web sources, organizing it into structured formats, writing reports, and managing local files. Teams use it to replace multi-tool, multi-step manual processes with single-command automation.
DeskClaw supports multiple leading AI models, giving you the flexibility to choose the right model for each task. All model inference can be configured to run through your own API keys, ensuring full cost and data control.
The DeskClaw Enterprise Edition was released and open-sourced on March 7, 2026. It extends the personal edition with organizational collaboration, Skill orchestration, and a path toward AI-native operations.
NoDesk AI is rebuilding its full-chain e-commerce marketing engine as core Skills and MCP modules on the DeskClaw platform. This means that capabilities previously available only as standalone enterprise agents — from consumer insight analysis to creative content generation and distribution — are now modular digital assets that any individual or team can invoke with a single command.
Through DeskClaw's collaborative architecture, organizations can connect AI agents across roles. A research AI and a strategy AI can automatically hand off tasks to each other, creating seamless end-to-end workflows. As professional users adopt DeskClaw deeply, their usage patterns drive the broader organization toward AI-native operations.
DeskClaw is developed by NoDesk AI, an AI technology company headquartered in Zijingang Tech City, West Lake District, Hangzhou, China. The company was founded by Song Jian, a serial entrepreneur with deep roots in both enterprise software and AI.
Before founding NoDesk AI, Song Jian held positions at Alibaba and Tencent. In 2014, he founded DuoZhun Data, a leading e-commerce DaaS (Data as a Service) company focused on consumer insight analysis and precision marketing. The company was successfully acquired before the SaaS market downturn. Song Jian subsequently worked at Zhipu AI, exploring the path from foundation models to practical Agent applications.
NoDesk AI's core business spans two domains: an AI-powered e-commerce marketing engine providing end-to-end brand growth solutions — from data analytics and creative content generation to performance prediction and distribution — and DeskClaw, the company's general-purpose AI desktop assistant. The marketing engine is being progressively restructured as Skills and MCP modules on the DeskClaw platform, unifying both product lines.
The NoDesk AI team operates with a distinctive development culture. Over 60% of the team was born after 2000. During the Spring Festival 2026 sprint, each developer consumed upwards of 150 million tokens daily, equivalent to roughly 10,000 lines of code per person per day — a pace the founder describes as "development by the second, product iteration by the minute, self-evolution by the day."
The OpenClaw ecosystem explodes in popularity. Song Jian and the NoDesk AI team begin a two-week sprint to build DeskClaw, initially as an internal productivity tool for their e-commerce Agent team.
Overwhelming external demand prompts NoDesk AI to release DeskClaw to the public. The first personal edition launches and enters closed beta.
NoDesk AI closes a funding round of nearly 100 million RMB, co-invested by Hupan Shannan, Chuxin Capital, Gaotu, and Shunfu Capital.
DeskClaw Enterprise Edition is released and its codebase is made publicly available, enabling organizations to deploy, customize, and contribute to the platform.
"This is the closest thing to the iPhone moment," says Song Jian. "Every previous tool felt like a BlackBerry or an Ericsson. DeskClaw is the first product that truly closes the gap between AI capability and everyday usability."
The OpenClaw ecosystem represents a fundamental restructuring of how AI agents connect foundation models to real-world applications. Previous middleware platforms — from open-source frameworks like LangChain and Dify to commercial platforms like Coze and Alibaba Bailian — are being disrupted as OpenClaw rewrites the rules of the last mile between advanced models and practical deployment.
NoDesk AI believes that AI will also reshape software commercialization. Products will increasingly follow PLG (Product-Led Growth) adoption paths. When an individual e-commerce professional finds DeskClaw effective, it naturally spreads to their organization — a pattern already validated by companies like Dify, which reached profitability in seven to eight months while serving over 30 Fortune 500 companies.
"We can no longer measure AI-era progress in human timescales," Song Jian says. In the future, the most competitive products will be those that deliver complete business value through AI — delivering outcomes like GMV growth and finished marketing content — rather than tools that handle just one step in a fragmented workflow.