OpenClaw goes viral worldwide: From Clawdbot to Moltbot, this AI agent is reshaping the era of "action-oriented artificial intelligence"

February 2 News, an open-source artificial intelligence agent called OpenClaw is rapidly spreading worldwide, expanding from Silicon Valley to the Chinese tech scene, and sparking intense discussions about whether “AI is moving toward autonomous decision-making.” The project was originally named Clawdbot, then renamed Moltbot, and finally gained popularity under the name OpenClaw, released by Austrian developer Peter Steinberger a few weeks ago.

The explosive attention to OpenClaw aligns perfectly with the current market’s high expectations for “autonomous task-completing AI agents.” Unlike traditional large models centered around conversation, these intelligent agents can perform real-world actions on behalf of users without continuous human instructions, and are seen by some industry insiders as “AI that can take action.”

OpenClaw runs directly on users’ operating systems and application layers, automatically handling tasks such as emails, calendars, web browsing, and online service interactions. Users need to deploy it locally or on a server and connect it to large models like Claude or ChatGPT. Although the installation process is not user-friendly for ordinary users, it has been integrated into communication platforms like WhatsApp, Telegram, and Discord, allowing remote control via text commands.

In real-world scenarios, users have used OpenClaw to automatically browse websites, generate PDF summaries, shop intelligently, manage schedules, and even send or delete emails on their behalf. Its “persistent memory” feature can store interaction history for several weeks, enabling it to gradually adapt to user habits and develop highly personalized operational logic.

As an open-source project, OpenClaw’s code is fully viewable and modifiable, which is a key reason for its rapid spread. The project has already received over 145,000 stars and 20,000 forks on GitHub. Initially tested extensively in the US tech scene, it later expanded into the Chinese market. Several Chinese cloud service and AI companies have begun integrating it into their own product ecosystems and supporting interaction with domestic models like DeepSeek.

Despite the high enthusiasm, controversy has also arisen. Some experts believe that OpenClaw’s deployment is complex, requires high computing power, and faces fierce competition from other intelligent agent products. Supporters argue that it can save a significant amount of repetitive work each week and is an important step toward general artificial intelligence.

IBM researcher Kaoutar El Maghraoui pointed out that this technology proves that intelligent agents are not exclusive to large corporations; once they gain system-level access, their capabilities far surpass traditional chatbots. However, security agencies hold a different view. Palo Alto Networks and Cisco warn that OpenClaw’s features—access to private data, contact with untrusted content, and communication with the outside world while retaining memory—pose a high risk of being manipulated to execute malicious commands or leak sensitive information.

Discussions around OpenClaw have also reignited due to Moltbook. Launched by Matt Schlicht, this platform allows AI agents to post, comment, and interact like users, even issuing declarations about “the end of the human era,” with some agents attempting to issue their own cryptocurrencies. In a related post retweeted by Elon Musk, Andrej Karpathy described it as “the closest real experience to a sci-fi scenario.”

Counterpoint Research analyst Marc Einstein stated that this viral spread is changing the public’s perception of intelligent agents, prompting people to reconsider the opportunities and risks they may bring. As projects like OpenClaw continue to evolve, an era where “everyone has their own private AI assistant” is accelerating toward reality.