Clawdbot: The Rise of an Autonomous AI Assistant and 72 Hours of Chaos
Clawdbot (now Moltbot) is a local AI assistant that captured global attention for several days. Highlighted as an agent capable of long-term memory and autonomous decision-making, it faced a series of incidents over 72 hours during its rebranding process, including account hijacking, fake token scams, and security vulnerabilities.
In January 2026, an open-source project emerged on GitHub that sparked tremendous buzz. Clawdbot, created by Austrian developer Peter Steinberger, wasn't just a simple chatbot—it was an autonomous AI agent capable of acting on behalf of users.
What is Clawdbot?
Clawdbot is a local AI assistant that runs directly on your computer (such as a Mac Mini). It integrates with over 50 messaging platforms including WhatsApp, Telegram, Slack, Discord, iMessage, and Signal, using AI models like Claude or GPT as its brain.
For example, if you send a WhatsApp message saying "Schedule a meeting for tomorrow at 3 PM," Clawdbot checks your calendar and creates the appointment. If you ask "Find the invoice I received last week," it searches your email and retrieves the file. You can even set it to send you daily weather and schedule updates proactively, without you having to ask.
With capabilities including web browsing, file management, and script execution, it goes beyond simple conversation to actually handle tasks like a real assistant.
Explosive Response
The developer community's reaction was explosive. Comments like "Finally, a real JARVIS" flooded in, and people started buying Mac Minis just to run Clawdbot. Tesla's AI Director Andrej Karpathy praised it, and MacStories called it "the future of personal AI assistants."
An AI with long-term memory, multi-platform capabilities, and autonomous decision-making—it was the kind of agent people had been waiting for.
Core Features
| Feature | Description |
|---|---|
| Long-term Memory | Maintains conversation context for days, weeks, or months |
| Autonomous Task Execution | Email responses, calendar management, script execution, web browsing |
| Multi-platform | Integrates with 50+ platforms including WhatsApp, Discord, Slack, iMessage, Telegram |
| Proactive Notifications | Sends reminders, briefings, and alerts without user requests |
| Local Execution | Runs directly on your computer, not in the cloud |
72 Hours of Chaos
But then came the twist. Just as popularity was soaring, a series of incidents unfolded over 72 hours.
First was the security issue. Security researchers warned that hundreds of Clawdbot servers were exposed to the internet. They were easily discoverable via Shodan searches, and some were completely open without authentication, exposing API keys, conversation logs, and login credentials. Google Cloud's VP directly warned people "do not install this."
Anthropic's Request and a 10-Second Mistake
To make matters worse, Anthropic requested a name change, citing that 'Clawd' was too similar to their 'Claude' trademark. The project was rebranded to Moltbot, inspired by how lobsters molt and grow new shells.
The problem arose during the rebranding process. While attempting to simultaneously change GitHub and X (Twitter) handles, crypto scammers hijacked the original accounts within 10 seconds. The scammers promoted fake tokens to tens of thousands of followers, and a fake $CLAWD token appeared on Solana, reaching a market cap of $16 million before crashing.
Unstable, But the Implications Are Clear
Clawdbot remains an unstable open-source project. Security vulnerabilities, rebranding chaos, scam damage—the risks are too high for immediate use.
However, what Clawdbot demonstrated is clear: the market's hunger for AI agents with long-term memory, autonomous decision-making, and multi-platform integration. The fact that a solo open-source developer implemented a 'practical agent' before AI frontier companies is significant.
What's certain is that the era of AI capable of long-term memory and autonomous decision-making is arriving.